146 results about "Weighted average method" patented technology

The invention relates to the technical field of wind field simulation, and provides a wind field numerical simulation method based on multiple meteorological data sources. The objective of the invention is to solve the problems of large simulation error, rough terrain boundary processing and poor turbulence model parameter adaptability caused by non-uniform coverage of a single data source, insufficient precision and unscientific multi-source fusion. The method is characterized by comprising the following steps: step 1, constructing a CFD three-dimensional grid based on a geometric model of a target area; 2, collecting original data of a ground station, satellite remote sensing, numerical forecasting and the like; 3, performing standardized preprocessing (abnormal value elimination, missing interpolation, radiation / geometric correction and resampling), determining a multi-source fusion weight by combining historical data analysis, and generating comprehensive meteorological data by adopting a weighted average method; and 4, inputting the comprehensive data into a CFD-RANS model, dynamically adjusting turbulence parameters, accurately setting terrain boundary conditions, and obtaining a wind field space-time distribution result through numerical solution. According to the invention, through combination of multi-source data fusion and CFD simulation, the wind field simulation precision and stability are improved.

The invention discloses a ship power grid transient electromagnetic interference grading evaluation method and system, relates to the technical field of ship power grid safety evaluation, realizes scientific grading evaluation of ship power grid transient peak interference, and provides a reliable basis for stable operation and equipment protection of a ship power grid. Comprising the following steps: measuring a transient peak sensitive threshold value of equipment by adopting a peak signal transduction sensitivity test; performing classification and importance coefficient assignment on the equipment according to the importance degree of the equipment; and calculating a weighted average sensitive threshold value of load equipment at the rear end of the distribution box by adopting a weighted average method. At the output end of a power distribution box of a ship power grid cabin, a transient peak conducted emission detection device is used for monitoring and collecting power grid transient peak conducted emission data in real time, and a power grid transient peak interference measured value is recognized from the power grid transient peak conducted emission data. Comparing the power grid transient peak interference measured value with the weighted average sensitive threshold value, and calculating a difference value between the two values; and according to a difference value result, carrying out ship power grid transient peak interference grading evaluation.

The invention provides a real-time monitoring method and system for a power transformer, and relates to the technical field of intelligent monitoring, and the method comprises the steps: carrying out the noise removal and time synchronization of original data through wavelet transform, and carrying out the recognition and elimination of abnormal data through an isolation forest algorithm, and obtaining the preprocessed data; key features are extracted from the preprocessed data, a weighted average method is adopted for fusion, and a comprehensive state index is generated; based on the comprehensive state index, using a fault classification model constructed based on a deep neural network to identify a fault type, using a long short-term memory network to analyze and predict the future state of the transformer, and calculating a transformer health score according to the comprehensive state index and the future state of the transformer; and according to the transformer health score, setting multi-level alarm thresholds for real-time alarm, and generating a maintenance suggestion in combination with the fault type. According to the invention, the accuracy of fault early warning is improved.

The invention provides a freight freight prediction method and system based on multi-dimensional feature fusion, and the method comprises the steps: constructing a multi-dimensional fusion feature based on an interaction relation between a multi-dimensional waybill feature in waybill data and a multi-dimensional external influence factor in external influence data; and obtaining a plurality of prediction sub-models based on multi-dimensional fusion feature training. And combining the plurality of prediction sub-models by adopting a weighted average method to obtain a prediction model. And under the condition that the query request is obtained, performing semantic analysis on the query request to obtain standardized query information, and inputting the query information into the prediction model to obtain freight freight prediction information corresponding to the query information. According to the scheme, the influence of external factors on freight prediction is introduced, the freight prediction accuracy is remarkably improved, the multiple prediction sub-models are combined according to the weight to obtain the prediction model, and the adaptability in different scenes is improved. The complete query information is obtained by performing semantic analysis on the query request, so that accurate execution of prediction can be guided, and the use experience of a user is improved.

The invention relates to the technical field of photovoltaic power generation and load prediction, and discloses a gateway electric quantity prediction method and system for a virtual power plant, and the method comprises the steps: carrying out the preprocessing of obtained historical multi-source weather, photovoltaic output and load power data; based on the preprocessed multi-source weather forecast data and the corresponding historical measured data, training a preset weather deviation correction model to correct irradiance and temperature in future weather forecast; fusing the corrected data through a weighted average method to construct meteorological input data; performing feature derivation processing on the meteorological input data and screening features strongly related to photovoltaic output and net load electricity consumption; a dynamic weighted integration model is further adopted to predict the photovoltaic power generation capacity and the net load power consumption; and determining the gateway electric quantity according to the difference value between the photovoltaic generating capacity predicted value and the net load electricity consumption predicted value. According to the method provided by the invention, the data quality, the meteorological input precision and the model prediction performance can be improved, and the generalization ability and the interpretability of the model are enhanced.

The invention discloses a detection and monitoring system for fault early warning in use of an insulating sling, relates to the technical field of electrical safety and intelligent monitoring, and is used for solving the problem that an internal progressive deterioration trend cannot be captured. A microwave resonance signal in the sling is obtained, the resonant frequency offset of each layer is extracted based on spectral decomposition, and a humidity distribution result is inversed and fed back in combination with a dielectric property database, and is used for setting a monitoring period and collecting operation data; a key area is screened based on a humidity diffusion coefficient, an evaluation area of which the risk level needs to be improved is identified by using a weighted average method, and the risk weight is calculated and the early warning time is set through a nonlinear model in combination with the physical attribute and the diffusion coefficient of the area, so that the early identification and dynamic early warning of the hidden fault of the sling are realized, and the safety and the monitoring precision are effectively improved. And the sudden failure risk is reduced.

The invention relates to the technical field of psychological evaluation, in particular to an AI model-based student negative emotion evaluation method, which comprises the following steps of: acquiring facial video data and mouse movement data of a student in real time; state data such as attention degree and hesitation degree of answering of the students are dynamically analyzed, and subsequent emotion test questions are dynamically adjusted according to an analysis result; after an emotion test question of a student is completed, the question weight is dynamically adjusted according to question difficulty, face video data and mouse movement data, the score of the test question is combined, and the emotion score of the student is quantified by adopting a weighted average method. And generating an emotion analysis report of the student according to the question difficulty, the face video data, the mouse movement data and the emotion score. By means of the mode, the problem that the accuracy of the emotion evaluation result of the student is low due to the fact that the face features of the tested student are analyzed in the testing process and data in multiple aspects such as the attention of the student is not considered can be solved.

The application discloses a kind of warship power grid transient electromagnetic interference grading evaluation method and system, it is related to warship power grid safety evaluation technical field, realize the scientific grading evaluation of warship power grid transient peak interference, provide reliable basis for the stable operation and equipment protection of warship power grid.It includes the following steps: using peak signal conduction sensitivity test, determine equipment transient peak sensitive threshold value.According to the importance of equipment, the importance coefficient of equipment is valued;Weighted average method is used to calculate the weighted average sensitive threshold value of the load equipment at the back end of distribution box.In the output end of the cabin distribution box of warship power grid, the transient peak conduction emission detection device is used to monitor and collect power grid transient peak conduction emission data in real time, and the power grid transient peak interference measurement value is identified from the data.The difference between the power grid transient peak interference measurement value and the weighted average sensitive threshold value is calculated;According to the difference result, the grading evaluation of warship power grid transient peak interference is carried out.

The invention discloses an audio and video generation method and system based on a large model, and relates to the technical field of artificial intelligence audio and video processing, and the method comprises the steps: collecting multi-modal data to extract features, and employing a pre-trained GLM-4 model to output a semantic consistency score; the method comprises the steps of defining an action space and a state space, calculating comprehensive rhythm density by using a weighted average method, calculating rewards of semantic consistency scores, constructing a fractional order Markov model, solving a rhythm phase difference by using a segmented Mitiger-Leveller function, defining a piecewise function PR, and solving the slope of the piecewise function PR by using an HHO optimization algorithm. The increment action of the audio and video is calculated, and correction is carried out by using an SBS-SMO sliding mode observer; fractional order Markov modeling is used, a segmented Mitiger-Leveller function is used for solving, the action space slope is regulated and controlled through an HHO optimization algorithm, and the semantic matching degree and rhythm coordination of audio and video generation are improved.

The application discloses a kind of rockfill dam deformation stability evaluation method based on multi-parameter analysis, it is related to water conservancy engineering technical field, and evaluation step is:S1, based on the data of monitoring point to rockfill dam is collected, data includes dam body deformation data and dam body contour data, data set is formed after data pre-processing;S2, dam body deformation data is converted into deformation time series according to time sequence, its fractal dimension is calculated by box counting method, and the mapping relationship between fractal dimension and rockfill dam deformation stability is established by combining historical data statistical analysis;S3, dam body contour data is imported into three-dimensional modeling software, and three-dimensional geometric model of rockfill dam is constructed.The application constructs comprehensive evaluation system, carries out correlation degree analysis and verification to fractal dimension, displacement strain index, obtains evaluation value based on weighted average method, intelligently divides stability grade, and provides intuitive decision basis for water conservancy engineering management.

The present application relates to the technical field of geological prediction, in particular to a tunnel rock mass quality identification method and system based on multi-scale drilling model fusion, specifically comprising the following steps: S1, collecting and preprocessing the field drilling data to obtain a test set; S2, inputting the test set into a multi-scale fusion model to obtain a tunnel rock mass quality identification result, the tunnel rock mass quality identification result being one of the prediction labels, and the prediction labels including relatively complete, relatively broken, broken and fissure; in the training process of the multi-scale fusion model, the field drilling data is collected, preprocessed and segmented in multiple scales, and the multi-scale model fusion is performed by using the error reciprocal weighted average method in the training process. The effect of weakening the shortcomings of each scale model, optimizing the prediction effect and enhancing the practicality of the technology is achieved.

The present application relates to the field of Bayesian networks, and discloses a social security case criminal behavior determination method, device, equipment and medium, which can construct an evidence element system and evidence structure attribute according to the characteristics of a social security case, and construct the correlation between the evidence and the behavior in the social security case. Based on the evidence element system, the evidence structure attribute and the correlation, a Bayesian evidence reasoning network of the social security case is constructed. The prior condition probability table of the Bayesian evidence reasoning network is generated based on the coordination coefficient method and the weighted average method. The evidence information obtained by investigating the scene of the social security case is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network determines the target behavior information and the corresponding occurrence probability according to the prior condition probability table and the evidence information. The present application can effectively reduce the consumption of human resources of case investigators and enhance the intelligent investigation level of social security cases.

The application discloses a molten silicon liquid level estimation method based on a hybrid adaptive resampling particle filter, and comprises the following steps: firstly, processing laser spot images collected by a CCD sensor to obtain observation data of the liquid level; then, establishing a system state equation of the molten silicon liquid level according to a kinematic principle to describe the movement of the liquid level; performing state estimation on a laser centroid vertical coordinate by using a particle filter algorithm, and obtaining a new particle set by using a hybrid adaptive resampling method; after re-normalizing all particle weights, outputting an optimal estimation value; and finally, smoothing the filtered data by using a moving weighted average method, and the smoothed result is the estimation value of the molten silicon liquid level. The application solves the problem that in the prior art, the variance of Gaussian variation is too large, particles jump out of a sampling range, and it is difficult to accurately estimate a real liquid level.

The invention relates to the technical field of artificial intelligence, and provides a multi-modal federated learning method based on a projection adaptation mechanism, and the method comprises the steps: loading a pre-trained cross-modal trunk model through a server, and completing the parameter initialization of a projection adaptation module; the server side distributes the global model to each client side, the client sides use local multi-modal data to carry out training, only projection adaptation module parameters are updated in the training process, cross-modal trunk model parameters are kept fixed, and meanwhile a symmetric cross entropy loss function is adopted to calculate a similarity matrix of text features and image features. And the client uploads the updated adaptive module parameters to the server side, and the server side aggregates the parameters in a weighted average mode to obtain an updated global adaptive module. The invention further discloses a system used for the method, the method and the system can still achieve the model accuracy equivalent to that of traditional centralized training in the complex environment that multiple clients hold heterogeneous data, and training efficiency, privacy security and performance expression are considered.

The invention relates to the technical field of building construction intelligent monitoring, and discloses a BIM-based bridge construction intelligent monitoring system, which comprises a comprehensive intelligent sensing module used for collecting various environmental data and bridge structure data in real time and carrying out preliminary data verification and error processing; the data fusion module is used for fusing various environmental data and bridge structure data by adopting a weighted average method; the intelligent decision support module is used for implementing natural language processing analysis based on the fused data and in combination with a construction site text record to extract an emotion score, generating a decision tree based on the emotion score and a set threshold value, and outputting a construction suggestion; a dynamic digital thread module; a cross-domain collaboration module; and a self-healing sensor network module. According to the method, the data fused by the weighted average method and the emotional score obtained through natural language processing and analysis act together, deep understanding of the construction site is achieved, the emotional state of the construction team can be analyzed, and specific construction suggestions can be generated according to the fused data.

The invention discloses a UWB single base station-based positioning method, system and device, a medium and a product, and relates to the technical field of positioning, and the method comprises the steps: comparing the position deviation of UWB initial positioning data and IMU positioning data with a set threshold value, and judging whether the UWB initial positioning data is abnormal or not; when the data is abnormal, performing strong compensation on the UWB initial positioning data by using IMU positioning data based on a Kalman filtering method; when the data is normal, performing weak compensation or non-compensation on the UWB initial positioning data by using IMU positioning data by adopting a weighted average method; when the UWB signal of the positioning tag is lost by the UWB single base station, the IMU positioning data is used as a positioning result, the UWB initial positioning data is corrected by using the IMU positioning data, the UWB single base station positioning data drift caused by factors such as multipath propagation and non-line-of-sight propagation can be effectively inhibited, and the positioning accuracy is improved.

The application discloses a kind of based on dual-camera width measurement's double-tower cylinder transport ovality on-line monitoring method, belong to wind power generation equipment state monitoring and intelligent operation and maintenance technical field.The method includes: after the side surface image of tower cylinder collected by camera is preprocessed, by edge extraction and least square method fitting, and further calculate the instantaneous projection pixel width;Scale factor and illumination compensation model are introduced, the actual physical width of the two sides of tower cylinder is obtained, and the weighted average method is used to obtain the short axis diameter of tower cylinder;According to the short axis diameter of tower cylinder and the ovality calculation model, the deformed long axis of ellipse is obtained, the real-time ovality of tower cylinder at present time is further calculated, and whether the real-time ovality meets the ovality upper limit requirement, the ovality change requirement and the ovality fluctuation amplitude requirement is judged, whether alarm signal is generated according to the determination result.The application method effectively solves the technical problem that tower cylinder is difficult to measure directly in double-tower cylinder parallel transport caused by tower cylinder shielding.

The application discloses a non-local mean-based adaptive SAR interferogram filtering method, and relates to the technical field of synthetic aperture radar. Firstly, two input SAR images are used to generate an interferogram and calculate the correlation coefficient of each pixel to form a correlation coefficient matrix. Phase information of the interferogram is extracted. Pixels that do not need filtering are screened. For pixels that need filtering, different search window radii are determined according to the correlation coefficient values, and the search window radius corresponding to the pixels that do not need filtering is set to 0. For each pixel that needs filtering, the weight of each pixel in the corresponding search window is determined, and the filtering result of each pixel is obtained by using a weighted average method to form a filtered SAR interferogram.

The invention relates to the technical field of artificial intelligence and big data, and discloses a vector compression reconstruction method for long text semantic retrieval, which comprises the following steps: firstly, segmenting a long text T into n text segments S1, S2,..., Sn according to semantic similarity or a fixed length; secondly, for each text segment Si, generating a local semantic vector li by using a BERT pre-trained language model, and converting the input text segment Si into a high-dimensional vector by the BERT model; then global vector generation is carried out, specifically, weighted average is carried out on the local vectors li, l2,..., ln of all the segments through a weighted average method, and a global semantic vector g is generated; carrying out explicit compression and key information storage; performing staged hybrid semantic retrieval; cross-segment semantic reconstruction and reasoning are carried out, and context information recovery is carried out; performing dynamic semantic reconstruction; and generating answers based on the reconstructed context. According to the method, deeper reasoning can be carried out, and more accurate and more comprehensive answers are generated.

The application discloses a mine shaft group drilling and blasting construction state monitoring method based on multi-source sensor fusion, and the method is as follows: sensors are arranged on a mine shaft group drilling and blasting construction site; a data acquisition and transmission module collects data of each sensor in real time and transmits the data to a data processing center; the data processing center removes outliers from original data and extracts feature information related to the drilling and blasting construction state; the feature information is subjected to data fusion processing, a comprehensive monitoring index is formed by using a weighted average method; the monitoring index is used to evaluate the construction state and to monitor the shaft drilling and blasting construction state in real time; the real-time monitoring result is fed back to a construction management system, used for adjusting and optimizing a construction plan, and the monitoring result is transmitted to construction personnel in the form of a chart so that the construction personnel can know the construction state at any time. The application improves the accuracy and reliability of the construction state evaluation.

The application discloses an intelligent mineral prediction method based on multi-element information fusion, and relates to the technical field of mineral exploration. The method comprises the following steps: obtaining a real data set by preprocessing geochemical data and remote sensing data; obtaining a simulation data set by data enhancement on the processed data; dividing the data set obtained by combining the real data set and the simulation data set into a training set and a test set; extracting detailed features from the data after feature extraction to obtain a first output feature map; extracting global features from the data after feature extraction to obtain a second output feature map; fusing the first output feature map and the second output feature map by a weighted average method to obtain an image fusion result; inputting the image fusion result into a feature decoder module to obtain a final fusion result; and inputting the final fusion result into a mineral prediction model based on a convolutional neural network to obtain a prediction result map. Thus, the fused image is used for mineral prediction, and the accuracy of mineral prediction is improved.

The invention provides a cross-modal fusion algorithm for fusing voice data and visual data, and relates to the technical field of multi-data fusion, and the algorithm specifically comprises the steps: S1, carrying out the feature extraction of the data information of voice data and visual image data; s2, generating a first risk factor and a second risk factor; s3, obtaining a comprehensive risk score, and constructing a corresponding weight distribution matrix; s4, generating a cross-modal consistency index used for representing dynamic uncertainty compensation; s5, acoustic dynamic weight distribution and optical multi-scale alignment control are executed; according to the invention, acoustic and optical output states are subjected to feedback coupling to form a closed loop optimization path, and a cooperative gain effect is realized; the contribution proportion of the acoustic data and the optical data to a cross-modal consistency index is determined through a weighted average method, and the uniformity and the performability of the acoustic data and the optical data in the physical parameter level are ensured through layered feature extraction and weight dynamic adjustment.

The present application relates to the field of construction engineering and geological monitoring technology, and discloses a vertical deformation monitoring method and system based on three-dimensional scanning and differential pressure level gauge. The method comprises the following steps: collecting point cloud data of a building structure in real time by using a three-dimensional scanner, and collecting vertical displacement data of the building structure in real time by using a differential pressure level gauge; obtaining the vertical coordinates of the point cloud by using a three-point registration algorithm, and calculating the vertical coordinate difference value of the point cloud to obtain a first difference value; calculating the average of the first vertical displacement data of the current monitoring period and the average of the second vertical displacement data of the previous monitoring period, and calculating the difference value between the average of the first vertical displacement data and the average of the second vertical displacement data to obtain a second difference value; and weighting the first difference value and the second difference value according to the standard deviation of the point cloud data and the standard deviation of the vertical displacement data, and calculating the final vertical deformation of the building structure by using a weighted average method. By implementing the present application, the cost of manual measurement is reduced, and the defect of single data source is changed.

The invention discloses an efficiency evaluation method and system based on air-ground integrated network analogue simulation confrontation, and relates to the technical field of network simulation and efficiency evaluation. The method comprises the following steps: constructing a space-ground integrated network topology simulation model by using a simulation engine; injecting a confrontation event, and simulating a confrontation scene; collecting network performance index data in the confrontation scene; on the basis of the collected index data, a weighted average method is adopted, and project weights included in all levels in a pre-constructed two-level performance evaluation system are utilized, a second-level score is firstly calculated, and then a first-level score is calculated on the basis of the second-level score; based on the score of the first level, utilizing a fuzzy evaluation method to obtain a fuzzy efficiency level of the space-ground integrated network; and taking the score of the first level and the fuzzy efficiency level as an efficiency evaluation result. According to the method, accurate reproduction of a multi-dimensional dynamic scene and closed-loop integration of simulation evaluation can be realized. And meanwhile, a multi-algorithm fusion method is adopted, so that the accuracy and robustness of evaluation are improved.

This invention discloses an intelligent oil pump oil quality monitoring method based on multi-source data fusion, belonging to the field of transformers. The method includes multi-source data acquisition, data processing, data fusion, result display, and long-distance data transmission: First, multi-dimensional raw data of the oil pump pipeline is collected through micro-water, pressure, and flow sensors and transmitted to a microcontroller control board; then, the raw data is denoised and filtered to obtain a standardized dataset; subsequently, a multi-source data fusion algorithm combining weighted average method and fuzzy logic is used to process the dataset, constructing an oil quality status assessment model to determine whether the oil quality has deteriorated; next, the monitoring results are visualized locally; finally, long-distance transmission of monitoring data and local-remote data synchronization are achieved. This invention improves the accuracy of oil quality monitoring through multi-parameter collaborative monitoring and data fusion, realizing local real-time display and remote real-time monitoring of oil quality status, adapting to the operational needs of transformer maintenance sites.

The invention discloses a bulk electron microscopic image splicing method based on a multi-scale image pyramid, and belongs to the technical field of image processing. According to the method, two-stage local abnormal value filtering and RANSAC enhancement based on local geometric constraints are introduced in the image registration stage, and image enhancement, feature re-extraction and multi-iteration TPS displacement field calculation are adopted in local deformation correction, so that the adaptability and robustness of the algorithm to a complex biological sample microstructure and potential deformation are jointly improved. By introducing a multi-scale image pyramid structure and a hierarchical processing strategy, a complex fusion problem of a high-resolution image is decomposed into a plurality of low-resolution sub-problems, and the calculation efficiency and the memory utilization rate are remarkably improved; and the problem of fuzzy texture details caused by a traditional weighted average method is effectively avoided.

The invention belongs to the technical field of battery charging, and provides a charging remaining time estimation method and device and a medium, and the method comprises the following steps: calculating the charging current of each temperature and SOC interval according to a first charging data table and the charging current under a preset charging rate; at each temperature, calculating a weighted average current value of the whole charging process by using the obtained charging current according to an SOC (State of Charge) interval through a weighted average method; calculating the basic charging time at each temperature according to the weighted average current value and the battery capacity; and determining the remaining charging time based on the basic charging time and the compensation time at the initial charging temperature. The method has the advantages that the charging time of other temperature points is estimated by using the charging MAP table in the battery parameters and the actual charging time at a certain temperature point, and the calculation result is more accurate while the workload is reduced.

The invention belongs to the technical field of underground water numerical simulation, and provides an underground water pollution migration simulation method and device, electronic equipment and a storage medium, and the method comprises the steps: constructing a three-dimensional octree grid, a control equation and a definite solution condition of a hydrogeological structure of a target region; carrying out discrete solution on the control equation in the three-dimensional octree grid by adopting a finite volume method taking a node as a center; constructing a post-processing approximation function and a local posterior error indicator on octree leaf subunits of the three-dimensional octree grid; determining a unit set contributed by a global error by adopting a Drfler marking strategy, then performing splitting or merging operation on leaf units, and performing grid updating by adopting a high-order interpolation or volume weighted average method; and according to a preset simulation termination condition, repeatedly executing underground water pollution migration simulation and finally performing visual display. According to the method, the simulation accuracy of water pollution migration is improved in a posterior error estimation mode.

The invention discloses a multi-model fusion wheat berry mildew screening method, equipment and a storage medium, relates to the technical field of food safety and detection, and solves the problem of how to quickly and accurately detect aflatoxin in wheat berries. A fluorescence signal in a wheat berry sample is excited through a laser, and laser is introduced into the wheat berry sample through an optical fiber; a fluorescence signal generated after excitation is transmitted to a spectrograph to collect fluorescence spectrum data in real time; preprocessing the fluorescence spectrum data through a combined strategy of an SG smoothing method and a standard normal variable transformation method SNV to eliminate noise and light scattering interference; and inputting the preprocessed fluorescence spectrum data into a plurality of trained machine learning models for prediction, fusing prediction results of the plurality of models through an improved weighted average method, and outputting the aflatoxin pollution level of the sample to be analyzed according to a multi-model fusion result. The detection time of the aflatoxin in the wheat berry is greatly shortened, and the detection efficiency is improved.

The invention relates to a method for evaluating the hand function state of a stroke patient, and belongs to the field of stroke, the method combines a Fugle-Meyer motion function evaluation scale and doctor suggestions, determines nine standard hand actions for evaluation, accurately obtains the three-dimensional coordinates of 21 key points of the hand of the patient by adopting a Mediape algorithm, provides an accurate data basis for evaluation, and improves the accuracy of the evaluation. According to the characteristics of each hand action, the geometric features with the highest distinguishing degree for healthy or sick states are selected as evaluation indexes, including angles and distances, and the hand function states of different rehabilitation levels can be stably distinguished. In the evaluation process, a weighted average method is adopted, reasonable weights are given to different fingers, the difference between healthy and affected hands is highlighted, the evaluation accuracy is further improved, meanwhile, a particle swarm optimization (PSO) algorithm is introduced to achieve automatic selection of a threshold value, subjectivity and uncertainty of manually setting the threshold value are avoided, and the evaluation accuracy is improved. And the evaluation result is more scientific and reasonable.