818 results about "Information gain" patented technology

The invention relates to the technical field of flood early warning, and discloses an urban flood disaster early warning method and system based on artificial intelligence, and the method comprises the steps: collecting five types of information, i.e., meteorological perception, hydrological monitoring, geographic space, urban operation and social perception in real time, and obtaining multi-source data with precise space-time coordinates; through preprocessing, gridding space-time alignment and key feature screening, rainfall accumulation and confluence evolution related features are extracted; constructing a physically constrained space-time fusion deep learning model, and outputting a future ponding depth prediction result in combination with a multi-head attention mechanism; environmental changes such as urban terrains and drainage facilities are adapted through incremental updating and transfer learning; and fusing the ponding depth, the influence range and the regional vulnerability characteristics to generate multi-level early warning, and synchronously outputting a spatial distribution map, a time evolution trend and affected object evaluation information. According to the invention, urban flood control and disaster reduction decision making and public accurate risk avoiding can be effectively supported.

The invention relates to the technical field of artificial intelligence and big data analysis, in particular to a data operation system and method based on a knowledge graph, and the method comprises the steps: extracting an entity and semantic relationship from multi-source business data, and constructing a dynamic evolvable initial knowledge graph; node features are aggregated, and multi-dimensional situation state vectors are generated in combination with gating loop unit modeling behavior path dependence; through a structure-semantic coupling attribution scoring mechanism, a statistical information gain and semantic similarity are fused to identify a core driving factor, and a causal regression model of the factor and an operation target is established; dynamically adjusting the edge weight and the structure of the atlas in real time, and triggering a new path discovery mechanism to continuously optimize the atlas; and according to a quantitative business target, reversely extracting a high-confidence influence path from the atlas, and generating a personalized strategy combination through intervention simulation and multi-target Pareto optimization, thereby realizing intelligent recommendation and decision closed loop driven by an operation target. According to the invention, higher-precision operation situation awareness and strategy generation are realized.

The invention provides a sewage system traceability analysis and intelligent monitoring method, system and device based on a graph neural network, and a storage medium, and belongs to the technical field of environment monitoring and artificial intelligence. The invention aims to solve the technical problems of low efficiency, low precision, difficulty in processing multi-source data, poor monitoring network and the like of the existing sewage system pollution tracing method. The method comprises the following steps: constructing a sewage system knowledge graph fusing multi-source heterogeneous data such as water quality and water volume; adopting a multi-scale graph neural network model to learn pollution propagation characteristics based on the knowledge graph; after a pollution event occurs, pollution path backtracking is carried out in combination with physical models such as flow conservation so as to identify a pollution source; bayesian inference is introduced to carry out uncertainty quantification on a traceability result so as to assess the credibility of the traceability result; and finally, dynamically optimizing the layout of the monitoring points based on information gain and other criteria. According to the invention, rapid and accurate positioning of the pollution source can be realized, and the method is suitable for intelligent supervision of an urban sewage system.

The invention relates to a data processing method and system for multi-source complex biological information data. According to the method, expression profile data, gene variation data and clinical survival data are collected, and unified standardization processing is carried out on the collected data. Feature alignment is performed on different source data based on sample identifiers, a joint feature expression matrix is constructed, and a context dependency relationship across data types is maintained. On the basis, multi-stage feature screening is carried out through Lasso regression and information gain evaluation in sequence, and an optimal feature subset used for modeling is obtained. And further training a risk scoring model by adopting a Cox proportional risk regression method, and calculating a risk scoring value of the sample by utilizing the constructed continuous scoring function. And finally, dividing the score value into a plurality of risk levels, and generating a survival curve of each level in combination with a Kaplan-Meier estimation method so as to verify the risk layering effect and prediction significance of the model. According to the method, the accuracy of biological information modeling can be improved, and the method has good universality and practical value.

The invention discloses a GIS disconnecting switch multi-state intelligent sensing system, and relates to the field of GIS disconnecting switch state monitoring. A self-calibration multi-mode sensor is deployed for multi-source data acquisition, and novel sensors including terahertz imaging and the like are included; in data preprocessing, deep learning noise reduction is applied, fuzzy entropy is used for dynamic weighted fusion, and abnormal values are processed by an improved algorithm; feature extraction is combined with a plurality of frontier algorithms to process vibration signals, and CRNN is used to analyze acoustic signals; the QPSO evidence theory is adopted for state fusion perception, the node relation is learned by means of GNN, and the weight is adjusted according to the information gain rate; state assessment and early warning are based on transfer learning, GAN and LSTM-attention mechanisms, and early warning priorities are ranked by FAHP. According to the invention, multi-mode accurate acquisition, intelligent data processing, deep feature mining, innovative fusion perception, accurate evaluation and early warning and efficient fault diagnosis and positioning are realized, the state of the GIS isolation switch can be comprehensively and accurately perceived, the system is self-learned and optimized, the equipment safety is guaranteed, the risk of a power system is reduced, and the power operation and maintenance benefits are improved.

The invention discloses a multi-source heterogeneous agricultural data element fusion system and method, and relates to the technical field of data fusion processing, and the method comprises the steps: carrying out the data screening of cleaned data; the screened data is preprocessed by the constructed agricultural field knowledge base, the logic consistency of the data is checked through a rule engine, and a corresponding coping strategy is adopted according to a check result; identifying abnormal points in the multi-source data by adopting a Gaussian mixture model and a constructed logic rule combination, and correcting or complementing the detected abnormal points in combination with a specific scene; key features are selected through information gain, multi-source fusion is carried out on the key features and original multi-source data, an abnormal mode library is established after abnormal data modes are classified, and source categories of the abnormal data are matched from a fault matching rule library according to abnormal features of the abnormal data; and the fault matching rule base is established, so that the source category of the abnormal data can be accurately identified, and a targeted solution is provided for processing the abnormal data.

The invention relates to the technical field of driver behavior recognition, in particular to an intelligent behavior monitoring and recognition system for coping with a vehicle driver, which comprises a multi-mode driver data information module used for acquiring first driver multi-dimensional state information in a preset historical driving period, extracting and generating a first dangerous behavior feature set and a first dangerous response feature set according to the multi-dimensional state information of the driver; the environment information acquisition module is used for acquiring first driving environment interaction risk information in a preset historical driving period; a risk assessment model construction module; a real-time data evaluation module; according to the invention, the dangerous behavior of the driver is identified by combining the multi-dimensional state information of the driver with the environmental information, and the parameters of the dangerous behavior risk assessment model are continuously updated through the real-time data, so that the accuracy of monitoring and identifying the behavior of the vehicle driver is improved; meanwhile, high-risk behaviors can be obtained in advance, and the driving risk is prevented.

The invention relates to the technical field of medical diagnosis processing, and discloses an intelligent disease diagnosis and differential diagnosis system based on a knowledge graph, and the system comprises a data input module which is used for receiving and standardizing clinical symptoms, signs, laboratory examination data and historical medical record data of a patient; the knowledge graph construction and updating module is used for constructing and updating a knowledge graph of diseases and symptoms according to the medical literature and the clinical data, and the knowledge graph automatically extracts an incidence relation between the symptoms and the diseases from the medical literature through a natural language processing technology; and the reasoning and diagnosis module is used for performing intelligent disease diagnosis and differential diagnosis. According to the method, by optimizing reasoning path selection and information gain calculation, the path with the most information content is selected from multiple reasoning paths for diagnosis reasoning, and the most representative path is selected by calculating the correlation degree between each symptom and the disease.

The invention relates to an arc fault detection method and system based on a dynamic fuzzy threshold and a storage medium, and the method comprises the steps: collecting the current circuit data of to-be-detected electrical equipment, and carrying out the feature extraction of the current circuit data, and obtaining a current feature parameter; the method comprises the following steps: training historical circuit data by taking a decision tree algorithm as a model architecture, and in the training process, performing parameter updating through a double-sliding window mechanism and optimizing a fuzzy threshold band range through an information gain maximization principle to obtain a dynamic fuzzy threshold decision model; and performing arc fault analysis on the current characteristic parameters through the dynamic fuzzy threshold decision model to obtain a fault detection result, and outputting early warning information when the fault detection result is that an arc fault occurs. According to the method, the dynamic fuzzy threshold value band is constructed through the Gaussian mixture model, and the self-adaptive adjustment of the threshold value is realized in combination with a double-sliding-window online updating mechanism. According to the technical scheme, the accuracy of arc faults can be remarkably improved, and particularly the false alarm rate is reduced.

The embodiment of the invention provides a water supply network pollution tracing and positioning method and device based on a graph theory. The method comprises the following steps: S1, abstracting a topological structure of the water supply network based on a hydraulic model of the water supply network, constructing an adjacent matrix of the water supply network based on the topological structure, and determining an initial dynamic detection point candidate range by combining the adjacent matrix of the water supply network and utilizing a breadth-first search algorithm in a graph theory; s2, according to the dynamic detection point candidate range, dynamic detection point candidate schemes are generated, the information gain of each dynamic detection point candidate scheme is evaluated, and a dynamic detection point most beneficial to pollution source positioning is selected based on the information gain; s3, repeatedly executing the step S2 to obtain water quality pollution information, and continuously narrowing the candidate range of the dynamic detection points; and S4, when the calculated node pollution source result meets the positioning success condition or the dynamic detection point candidate range is empty, determining the position of the pollution source. In this way, the position of the pollution source can be quickly positioned.

The invention relates to the technical field of unmanned aerial vehicle autonomous control, and discloses a power transmission line unmanned aerial vehicle autonomous inspection path planning method and a quality inspection system, and the method comprises the following steps: obtaining or constructing and dynamically evolving a cognitive digital twin containing geometric, semantic and uncertainty information; based on the first task intention, aiming at reducing the uncertainty of the cognitive digital twinning, generating a first flight path by optimizing information gain; controlling the unmanned aerial vehicle to fly along the first flight path, and performing in-service quality inspection on the acquired data to obtain a quantized target component defect suspected degree; and when the defect suspected degree meets a preset risk triggering condition, dynamically generating a second task intention with a higher priority, and generating a second flight path for fine detailed investigation based on the intention so as to replace or modify the first flight path. According to the invention, the unmanned aerial vehicle is converted into an active diagnosis agent from a passive path executor, and the autonomy, reliability and efficiency of routing inspection are significantly improved.

The invention discloses a remote sensing image multi-source heterogeneous data fusion processing method and system, and the method comprises the steps: extracting global information from remote sensing image data through employing a convolutional neural network, extracting image local features through cutting operation, and capturing local feature information in the remote sensing image data; constructing a cross-time-domain attention mechanism for the local feature information through a cyclic matrix to extract mutual information among different modal variables, and screening highly-associated cross-time-domain key information; establishing a cross-time-domain sensing hierarchical aggregation module for the cross-time-domain key information, and obtaining detail information and edge information of the remote sensing image; acquiring global fusion data by adopting an asymptotic fusion strategy, and introducing a loss function to reduce a semantic gap; according to the method, missing information is repaired by adopting an interactive network model of mixed contrast learning, complete real-time remote sensing image multi-source heterogeneous fusion data is obtained, and efficient and high-precision fusion processing of the multi-source remote sensing data is realized by constructing a multi-collaborative deep fusion framework.

The invention provides a landslide risk dynamic early warning method and system based on a Bayesian network, and belongs to the field of geological disaster intelligent prediction, and the method comprises the steps: constructing a disaster-inducing factor data set, and screening key factors based on a Pearson's correlation coefficient and an information gain method; an FP-Growth algorithm is further utilized to extract association rules among high-confidence factors, a Bayesian network structure is guided to be optimized, and a Bayesian network model with a causal relationship is constructed; the model supports an incremental learning mechanism based on a newly added landslide sample, can dynamically update a conditional probability table, and realizes landslide probability prediction and risk grade division in combination with a Bayesian forward reasoning result. The method has the advantages of being high in causal reasoning ability, high in model structure expression ability, excellent in prediction precision and capable of supporting real-time updating and risk partition, and is suitable for an intelligent risk assessment and early warning system for landslide disasters.

The invention provides a prefabricated dish semi-finished product defect identification method and system with an AI algorithm, and the method comprises the steps: extracting semantic features from a refined defect candidate region set, carrying out the feature mapping of each region through a deep convolutional network according to the demands of atypical defect identification, and obtaining the defect description represented by a high-dimensional feature vector; marking original image data through a final defect identification result, and for inhibition of complex background interference, performing outward expansion from defect edge features by adopting a region growing algorithm to obtain complete defect region boundary information; after complete defect area boundary information is obtained, defect distribution changes of continuous batches of images are compared through a time sequence analysis method according to the monitoring requirement of production process fluctuation, and the quantitative basis of process adjustment is determined.

The invention relates to the field of detection and calibration of numerical control machining equipment, and discloses a quick calibration method for a CNC machine tool cutter bit, which comprises the following steps: S1, collecting state data in the operation process of the cutter bit; s2, carrying out preprocessing on the state data; s3, inputting the preprocessed state data into a tool wear identification model, and outputting tool wear state information; s4, acquiring spatial point cloud data of the end part of the tool bit, and calculating spatial error information between the point cloud and a preset tool reference model; and S5, calculating the position offset and the attitude error of the tool bit in the three-dimensional space based on the space error information. By collecting the physical state data in the operation process of the tool bit and combining with the wear recognition model of the time sequence modeling mechanism, the wear state of the tool can be accurately recognized in real time, so that the problem of unstable recognition precision caused by environmental interference in a traditional method is solved, and the recognition accuracy and real-time performance are remarkably improved.

The invention provides an industrial data quality evaluation method based on micro-service, a medium and a system, and belongs to the technical field of electrical digital data process.The industrial data quality evaluation method comprises the steps that firstly, multi-source heterogeneous data is collected through a distributed network, and an evaluation system containing deviation rate, completeness rate and timeliness indexes is established; extracting features by using a deep learning model, and generating a feature probability distribution matrix; then, constructing a state transition matrix based on the characteristic entropy and the fluctuation coefficient, deploying an evaluation unit through the micro-service architecture, and establishing a transverse and longitudinal evaluation network to calculate information gain and association strength; and finally, performing feature fusion through an improved pyramid structure, realizing quality scoring by adopting a three-layer evaluation equation set, ensuring scientificity of a scoring result through logarithm mapping, and realizing dynamic evaluation of industrial data quality. The technical problem that in the prior art, an industrial data quality evaluation method is difficult to adapt to dynamic changes of multi-source heterogeneous data is solved.

The invention relates to the technical field of power equipment fault diagnosis, and discloses a power plant equipment multistage fault diagnosis method and system based on a dynamic decision tree, and the method comprises the steps: 1, collecting equipment operation data through multiple sensors, and carrying out the preprocessing of multi-source data; comprising noise filtering, outlier elimination and missing value filling; step 2, extracting feature values according to the time domain features and the frequency domain features, constructing a state decision tree model based on a C4.5 algorithm, optimizing attribute split points by an information gain rate, and optimizing generalization ability by an REP post pruning strategy; and step 3, monitoring operation data of the power plant equipment based on the state decision tree model, diagnosing and predicting equipment faults according to the monitoring data, outputting fault levels according to monitoring results, and triggering corresponding grading responses. The fuel power plant equipment fault diagnosis method provided by the invention has the advantages of high precision, low false alarm, capability of effectively distinguishing fault levels and dynamic self-adaptive capability.

The invention provides a radar anti-interference strategy generation method based on reinforcement learning by introducing a reinforcement learning technology into a radar anti-interference strategy generation problem on the basis that a radar anti-interference decision generation process has an obvious Markov decision property. And an optimization solution process generated by the radar anti-interference strategy is converted into an optimization solution process of reinforcement learning. According to the invention, reinforcement learning is introduced, and mechanisms such as a dynamic learning rate, action exploration, advantage normalization, reward scaling and hyperbolic tangent activation are combined, so that the radar anti-interference strategy generation method for near-end strategy optimization of asynchronous advantage actors and commentators and normalized scaling tangent based on dynamic learning exploration is provided. According to the method, high-dimensional features of different types of interferences can be accurately extracted in the time-frequency domain, the precision of anti-interference pattern selection and parameter selection and the environmental adaptability are improved, and thus the anti-interference capability and the information acquisition capability of the radar are improved.

The invention belongs to the technical field of wireless positioning, and discloses a non-line-of-sight interference-resistant dynamic positioning method and a base station layout method for positioning. Comprising the following steps: calculating original distance measurement results of each initial base station in a label and a sub-region, obtaining current shielding probability and current signal intensity of the initial base station and current geometric accuracy factors of different initial base station combinations in an initial base station set, selecting candidate base stations, calculating information gain of each candidate base station pair, and obtaining the information gain of each candidate base station pair; selecting candidate base stations with information gains meeting preset requirements to construct a target base station subset, inputting an original ranging result of the target base station subset into a Kalman filter based on a motion state space model for processing, and compensating an NLOS time delay error through a sliding window mean value to obtain a positioning result of a label and an NLOS time delay error estimation value, the problem that the positioning result is not accurate enough in a complex environment is solved.

The invention relates to the technical field of circuit assembly type selection, and particularly discloses an optimized matching circuit assembly rapid type selection method and system, which are applied to a circuit assembly characteristic database, and comprises the following steps: obtaining electrical parameter information corresponding to a plurality of candidate circuit assemblies; a virtual test model is constructed based on the electrical parameter information, virtual circuit operation is carried out based on the virtual test model, virtual operation information is generated, and the virtual operation information comprises a plurality of virtual dynamic operation parameters corresponding to a plurality of preset test time points; and obtaining operation performance information of virtual circuit operation according to the virtual operation information. According to the method, in the evaluation link, the voltage, temperature, power and other parameters are subjected to normalization processing, the multi-dimensional correlation change parameters are generated, the problem that traditional model selection neglects the component performance correlation relation is solved, the comprehensive performance of the component in a complex scene can be accurately evaluated, and a comprehensive basis is provided for model selection.

The invention discloses a knowledge graph diffusion recommendation method based on information gain guidance, and the method comprises the steps: carrying out the preprocessing operation of an original knowledge graph and user-article interaction data, calculating information gain, and obtaining a user feature weight matrix; constructing a knowledge graph denoising process based on a diffusion model based on the user feature weight matrix, and training the diffusion model by using an original knowledge graph and user-article interaction data; training a recommendation model by using the de-noised knowledge spectrogram, heterogeneous knowledge aggregation and contrast reinforcement learning to obtain a trained recommendation model; and inputting a to-be-predicted user ID and user-article interaction data into the trained recommendation model to obtain a related recommendation list of the user. By combining the knowledge graph diffusion model, the heterogeneous knowledge aggregation mechanism and the comparative learning strategy, the performance of the recommendation system is remarkably improved.

The invention provides an intelligent medical multi-round dialogue diagnosis reasoning method and system based on deep learning, and relates to the technical field of deep learning, and the method comprises the steps: constructing time sequence features through employing an attention mechanism for time sequence symptom description, and carrying out bidirectional sequence modeling to generate comprehensive symptom features; using the medical knowledge graph to detect logic contradictions and information loss to generate standardized features; calculating information gain of an inquiry direction based on a deep neural network to generate optimal inquiry content; and iteratively updating according to user feedback until the diagnosis information entropy is lower than a threshold value, and outputting a diagnosis result. According to the invention, the accuracy and efficiency of medical diagnosis are improved.

The invention provides a multi-agent collaborative exploration method for constructing an internal individual reward based on information gain. The method comprises the following steps: acquiring state information and action information of agents, creating a multi-agent reinforcement learning environment, and initializing parameters of a dynamic model; constructing a CVAE as a dynamic model, compressing a state space to a potential space, and predicting state distribution; optimizing model parameters by using reconstruction loss and KL divergence loss to obtain a trained CVAE model; estimating a priori entropy and a posteriori entropy of agent state transition through the trained model, and calculating an information gain; designing an exploration frequency factor, and calculating an internal individual reward in combination with information gain; and according to the internal individual rewards and the external rewards, generating a utilization strategy and an exploration strategy of the intelligent agent for collaborative exploration, and guiding the intelligent agent to select an action to be executed to complete a task target. The method provided by the invention is helpful for the intelligent agent to explore an unknown area, and the exploration capability and the task completion rate of the intelligent agent are remarkably improved.

The invention discloses a water quality evaluation model construction method based on a fruit fly-whale collaborative optimization algorithm, and aims to solve the problems that water resource pollution is serious, and a traditional water quality evaluation method is long in time consumption, high in cost and difficult to monitor in real time. Key water quality parameter characteristics are determined through information gain, and a sample data set is constructed and divided. Thirdly, establishing a support vector machine model, and setting a penalty coefficient and a kernel function parameter as a to-be-optimized combination; and searching an optimal solution globally by using a fruit fly optimization algorithm, taking the optimal solution as an initial solution of the whale algorithm, and further optimizing to obtain a parameter combination optimal solution. And configuring a support vector machine by using the optimal solution to construct an original model, and training and testing to obtain a final water quality evaluation model. The model is high in accuracy, can quickly process a large number of samples, is easy to integrate to an existing monitoring platform, and provides an efficient and accurate new way for water quality evaluation.

The invention relates to the technical field of communication, and discloses a three-dimensional radio environment map construction method and system, a terminal and a storage medium, and the core is to construct a'base station unmanned aerial vehicle 'bidirectional interaction closed-loop optimization framework to realize efficient construction of a high-precision map. Self-adaptively fusing sparse radio measurement data and environmental building structure features; introducing a confidence evaluation mechanism based on adversarial learning and position weighting, and generating a pixel-by-pixel confidence map; an intelligent planning method based on a trajectory diffusion model is designed, local perception constraint and long-term information gain are cooperated with a classifier-free guide mechanism, and an optimal trajectory considering both safety and sampling efficiency is generated; and a continuously self-optimized closed-loop system is formed through newly acquired data of the unmanned aerial vehicle and periodical updating of the model. According to the invention, a high-reliability technical basis is provided for applications such as urban air communication and spectrum resource management.

The invention provides a dynamic lung compliance monitoring method based on electrical impedance tomography, and the method comprises the steps: obtaining an electrical impedance signal and pressure-volume curve data, recording a time sequence signal of lung tissue in a dynamic ventilation process through a synchronous collection device, and obtaining an original multi-dimensional data set; aiming at the second feature set, eliminating redundant parameters by applying a feature screening mechanism, and evaluating the importance of each feature in different breathing stages through an information gain algorithm to obtain a third feature set; constructing a dynamic compliance monitoring model through the weighted feature set, and fitting a lung compliance change trend by using a support vector regression algorithm to obtain a real-time prediction result; and according to the updated prediction result, evaluating the improvement degree of the calculation efficiency, and judging whether the real-time performance meets clinical requirements or not by recording the model operation time and the resource occupancy rate so as to obtain final monitoring output.

The invention belongs to the technical field of artificial intelligence and medical information, and discloses a multi-round inquiry method and system based on a large language model and session state tracking. According to the method, extraction and synonym normalization are carried out for key medical elements, and high-confidence filling and conflict resolution are continuously completed in multiple rounds of conversations; and fusing the red flag symptom rule and model prediction, and carrying out hierarchical scoring and security constraint generation on individual risks. The information gain maximization serves as a target, and the next round of clarification problem is generated in a self-adaptive mode under the risk constraint; and through cooperation of a large language model and a knowledge base / knowledge graph, sorting and gate type calibration are carried out on candidate diseases and matched departments, and doctor-seeing suggestions, examination suggestions and medication precautions are generated. Finally, efficient understanding and multi-round reasoning of the unstructured symptom information are realized through joint supervision of the session state, the slot confidence and the risk hierarchy.

According to the traffic flow feature processing method and device provided by the embodiment of the invention, a multi-source data preprocessing mechanism is innovatively constructed, and the data quality is effectively improved through time sequence interpolation and abnormal value detection. An intelligent feature extraction model based on a random forest is designed, and high-dimensional sparse features are converted into low-dimensional dense representation in combination with information gain splitting and principal component analysis. A multi-level feature fusion mechanism is introduced, and accurate description of traffic flow features is realized through self-attention calculation and residual optimization. According to the method, the defects of the traditional technology in the aspects of data processing, feature extraction, fusion expression and the like are effectively overcome, and the accuracy and reliability of traffic flow feature processing are remarkably improved.

The invention discloses an anti-shielding three-dimensional scanning measurement field construction method for large-size measurement, and belongs to the technical field of optical three-dimensional measurement. The method comprises the following steps: firstly, performing adaptive sampling based on geometric features of a workpiece to generate a candidate mark point set; secondly, constructing a virtual measurement scene, simulating a dynamic scanning process by using a ray tracing technology, and calculating the visibility and observation quality of each point under multiple viewpoints; then measuring precision is quantified by using a Fisher information matrix, and a layout optimization model which takes maximization of information gain as a target and meets coverage rate and engineering constraint at the same time is established; and finally, solving the model to obtain an optimal mark point subset, and verifying and outputting the optimal mark point subset. According to the method, the problems of measurement interruption, low point distribution efficiency and difficulty in guaranteeing precision caused by shielding of a large-size complex component in scanning are solved, and automatic construction of a high-robustness and high-precision measurement field is realized.

The invention relates to a dynamic gesture recognition method for a multi-feature fusion network under complex scene interference, and the method comprises the steps: obtaining a baseband intermediate frequency signal, carrying out the processing of the baseband intermediate frequency signal, and obtaining a distance-time diagram and a Doppler-time diagram; inputting the distance-time graph and the Doppler-time graph into a gesture recognition model to obtain a gesture recognition result; the gesture recognition model is obtained by training a lightweight neural network model by using a training set; and extracting space displacement features and gesture speed change information of the distance-time chart and the Doppler-time chart based on a DSCC module in the gesture recognition model, obtaining a feature map, combining the feature map with position codes, inputting the feature map into a Transform module for global feature extraction, and obtaining a gesture recognition result. According to the method, through extraction and fusion of multiple features, the distance, speed and time information of the target gesture is expressed more comprehensively, and the richness and expression ability of the gesture features in a complex scene are improved.