119 results about "Local linear" patented technology

The invention provides a network security monitoring method and system based on distributed nodes, and the method comprises the steps: obtaining an inter-node interaction record of each node in a distributed network in a preset communication period and the identification information in a protocol interaction process, and obtaining a node communication data set; the method comprises the following steps: constructing a multi-dimensional phase space reconstruction matrix containing communication time sequence association features and protocol identifier association features, carrying out nonlinear dynamic feature analysis on the matrix, extracting a chaotic feature value set of node communication behaviors, constructing a topological association structure of node communication features through a local linear relationship, mapping the set to a topological space of a preset dimension, and constructing a topological structure of the node communication behaviors. And generating an attack mode topology expression vector, and finally analyzing an abnormal state propagation process in the distributed network through an inter-node energy propagation rule based on the attack mode topology expression vector to obtain a network security detection result. According to the invention, the accuracy and dynamic analysis capability of distributed network security monitoring can be effectively improved.

The invention belongs to the technical field of image processing, and particularly relates to an oil and gas engineering supporting facility intelligent detection method based on machine vision, and the method comprises the steps: constructing a structure tensor matrix based on the gradient of pixel points, calculating a local linear structure response value according to two feature values of the structure tensor matrix, and calculating a local linear structure response value; according to the direction alignment degree of the feature vectors of the pixel points on the local path, the crack continuity of the pixel points is calculated, the morphological significance values of the pixel points are comprehensively obtained and used for adjusting basic contrast parameters, anisotropic diffusion processing is carried out on the to-be-detected image of the oil and gas facility based on the obtained self-adaptive contrast parameters, and the to-be-detected image of the oil and gas facility is obtained. An enhanced image is obtained through multiple times of iterative updating, and a defect detection result graph is obtained through edge detection. According to the method, the problem that false defects are easily misjudged as cracks in a traditional method is solved, and the detection accuracy and reliability are remarkably improved.

The invention discloses a new engineering course teaching evaluation method based on a knowledge-ability-quality triple atlas, and belongs to the technical field of intelligent education and intelligent control. Collecting data of three dimensions of knowledge, ability and quality of students to construct a unified state vector, and describing dynamic evolution of a learning process by combining a state updating and prediction model of a residual network; a multi-cell filtering method is adopted to carry out state set estimation, linear propagation and residual local linearization are combined in the prediction step, and a Lipschitz upper bound is utilized to carry out external connection on a nonlinear residual to ensure the safety and credibility of state interval estimation; in the updating step, set tightening is achieved through prediction strip intersection and generator contraction, coverage rate calibration based on quantiles is introduced, and the confidence level of set estimation is ensured. According to the method, the learning state of the student can be dynamically, stably and interpretably estimated, accurate and efficient course adjustment optimization is realized, and the method has a good application prospect and popularization value.

The invention relates to the technical field of image data processing, in particular to an induction furnace coil temperature field visual monitoring method based on thermal imaging, and the method comprises the steps: obtaining the three-dimensional temperature field data of an induction furnace coil, and extracting high-temperature voxels; an adaptive weight is constructed for each high-temperature voxel, and the weight is composed of two parts: 1, the local linearity of the voxel is determined by calculating the projection density of the voxel in multiple directions; secondly, obtaining a colinearity score by calculating the consistency of the candidate voxel and other candidate voxels in the neighborhood of the candidate voxel in the local main direction; and performing weighted Hough transform on the candidate voxels to vote by taking the product of the two as an adaptive weight. Through the intelligent weighting mechanism, real linear fault signals can be effectively amplified, noise interference is suppressed, and the detection accuracy and robustness of early discontinuous linear high-temperature faults are improved.

The invention relates to the technical field of dam safety monitoring, and discloses a multi-source remote sensing data deep learning dam slope monitoring method, which comprises the following steps: resampling each mode to a common ground grid, and calculating robust statistics and a time stability agent on grid and block scales; determining a reference image according to the block-level robust score, and adaptively setting a local displacement search range with a high-intensity centroid difference; evaluating the discrete candidate displacement in a grid neighborhood by using a median absolute difference to obtain a local displacement field and a residual error; three types of subitems are constructed based on robust noise, time stability and registration residual errors, and pixel-level reliability weights are adaptively synthesized through logarithmic variance proportions among modes; analyzing and solving local linear mapping in a neighborhood by using a weight-weighted observation matrix, and calculating a weighted residual error according to the local linear mapping; a binary and probability anomaly graph is generated with a robust threshold.

The invention discloses a transformer abnormity identification method based on voiceprint feature analysis, and belongs to the field of power equipment state monitoring and intelligent diagnosis. The method comprises the following steps: firstly, analyzing an iron core acoustic mechanism based on a magnetostrictive effect, and establishing a three-dimensional model through finite element simulation to obtain vibration and sound field characteristics; in a complex substation environment, a hybrid noise reduction method combining density peak clustering and a CEEMDAN-wavelet threshold is provided, and the signal-to-noise ratio is effectively improved. Then extracting Mel-frequency cepstrum coefficients (MFCC) and spectrum features, and performing local linear embedding (LLE) dimension reduction to form a compact feature set; in the recognition stage, a convolutional neural network framework is designed, specifically, a spectrogram and an energy spectrum are modeled through a two-dimensional CNN, an MFCC tensor obtained after dimensionality reduction is modeled through a three-dimensional CNN, and accurate diagnosis of mechanical faults such as core looseness is achieved. The method has the advantages of being non-contact, anti-noise and high in recognition precision, and real-time diagnosis and early warning of mechanical abnormity of the transformer can be achieved under complex working conditions.

The invention relates to the field of fan spindle state monitoring, and aims at synchronously acquiring sound and vibration signals through multiple channels, realizing nanosecond time alignment by adopting a precision time protocol, and performing denoising and normalization preprocessing on the signals to improve data integration and signal fidelity. Furthermore, short-time Fourier transform and continuous wavelet transform are combined to extract multi-scale time-frequency features, and a high-dimensional combined feature vector is generated in combination with cross-correlation analysis. And mapping the feature vectors to a low-dimensional manifold space through a local linear embedding algorithm, and constructing a dynamic mode reference template. Indexes such as curvature, track length and direction entropy are monitored in real time, whether the spindle has an abnormal evolution trend or not is judged through a self-adaptive curvature threshold, and abnormity judgment is achieved in combination with track backtracking verification. According to the scheme, the abnormal starting boundary of the spindle state can be caught in a refined mode, and the early warning and operation and maintenance response capacity of fan operation is effectively improved.

The invention relates to the technical field of computer-aided engineering, and discloses an intelligent water conservancy design simulation system based on digital twinning, which comprises a parameterized geometric feature extraction module, a reduced-order model database, a tangent space linear deduction module and a flow field reconstruction module. According to the method, nonlinear flow field solution is dimensionally reduced into deterministic matrix vector multiplication, traditional statistical regression and iterative solution are replaced, it is ensured that a reconstruction result evolves in the tangential direction of a physical gradient, non-physical guess of a statistical model in a sparse sample area is avoided, and the calculation accuracy is improved. And high-fidelity flow field reconstruction with real-time performance and physical evolution conservation is realized.

The invention relates to a dynamic light beam shaping system and method based on a spatial light modulator. The system comprises a laser light source module, a beam expanding collimation light path, a light beam modulator, a focusing optical assembly, a detection and feedback module and a control and optimization module. According to the method, the uniformity of the output flat-topped beam can be corrected by virtue of a random parallel gradient descent (SPGD) algorithm and combining a Zernike polynomial as a control variable. Meanwhile, a composite performance index comprehensively considering energy distribution and shape features is constructed, the correction process of the flat-topped beam is accelerated by adopting a mode of gradually improving Zernike polynomial orders in multiple stages, the convergence trend of the performance index is used as a judgment basis of stage switching, stage independent variable dimension expansion is implemented, and the correction precision of the flat-topped beam is improved. And the convergence speed is obviously improved. Besides, adaptive selection of disturbance amplitude and learning rate parameters in the SPGD algorithm is realized by using noise estimation and a local linearity index, and the robustness of the system to noise and environmental disturbance is enhanced.

The invention belongs to the technical field of autonomous underwater vehicle numerical simulation, and discloses an AUV physical field prediction method based on manifold learning and deep learning, comprising the following steps: performing nonlinear dimension reduction on high-dimensional simulation physical field data based on an Isomap manifold learning algorithm, and extracting low-dimensional manifold features; constructing a deep neural network (DNN) model, and establishing a mapping relation between the working condition parameters and the low-dimensional manifold features; inputting new working condition parameters to the trained deep neural network model, and predicting corresponding low-dimensional manifold features; isomap inverse mapping is realized based on a local linear embedding algorithm, and predicted low-dimensional manifold features are reconstructed into high-dimensional physical field data. Through collaborative prediction of manifold learning and deep learning, the problem that a traditional method is low in calculation efficiency in high-dimensional physical field prediction is solved, prediction precision and real-time performance are remarkably improved, and an efficient tool is provided for AUV design optimization and dynamic control.

The invention provides a lithium battery cargo abnormal temperature rise identification method, and belongs to the technical field of lithium battery customs detection.The lithium battery cargo abnormal temperature rise identification method comprises the steps that a temperature sensor array is arranged on the surface of a lithium battery cargo stacking body to collect multi-point temperature time sequence data, and trend characteristics are extracted through wavelet packet decomposition denoising and singular spectrum analysis; establishing a state space model based on a heat conduction partial differential equation, estimating an internal temperature field state by using Kalman filtering, solving a heat conduction inverse problem by using a conjugate gradient regularization iterative algorithm to invert an internal three-dimensional temperature distribution field, and inputting an inversion result and statistical characteristics into a thermal anomaly identification model fused with manifold learning. Dimensionality reduction is carried out through a local linear embedding algorithm, a mahalanobis distance is calculated in a low-dimensional manifold space, an abnormal temperature rise risk score is output, when the score exceeds a preset threshold value, early warning is triggered, and the technical problem that the abnormal temperature rise in the lithium battery cargo stacking body is difficult to accurately recognize through surface temperature measurement is solved.

The invention provides a model-free prediction energy storage converter voltage control method based on a Kalman observer, and the method is characterized in that the method comprises the following specific steps: 1, constructing a super-local model of a dual-active bridge converter, and simplifying the system description through local linearization; 2, designing a Kalman observer to estimate and compensate lumped disturbance in real time, and combining dynamic matrix updating and gain adaptive adjustment; 3, designing a cost function to solve an optimal shift ratio, and realizing model-free prediction voltage control; according to the control method provided by the invention, the robustness of the converter under the conditions of model mismatch and working condition change is effectively improved, and meanwhile, a good dynamic response characteristic can be kept.

The invention belongs to the technical field of model simulation, and particularly relates to an efficient dynamics simulation analysis method based on a Fourier neural operator. The present invention proposes FNO-Speed, and a series of comprehensive solutions for inefficient operations that the FNO solver does not fully utilize hardware. According to the method, two unique optimization methods are adopted, and comprise a multi-level parallel implicit image-to-column general matrix multiplication optimization strategy and a user-defined size high-frequency signal filtering algorithm. According to the method, efficient general matrix multiplication is achieved through an implicit image-to-column and data division strategy to replace pointwise convolution, and fragmentary calculation of frequency domain local linear transformation is eliminated through the latter. The FNO-Speed makes full use of the memory bandwidth, improves the calculation efficiency, and aims to solve the problems of low utilization rate of calculation resources and delay influence caused by large-scale data access calculation.

The invention discloses a multi-angle laser radar vertical two-dimensional wind field inversion method. The method comprises the following steps: firstly, setting a measurement center and sampling parameters of a measurement area; acquiring multi-angle radial wind speed data; introducing a multi-point local sampling strategy to each inversion center point, and establishing a local multi-constraint inversion equation set by selecting a plurality of observation points close to the spatial position of the center point and combining the observation angle and radial wind speed information; according to the method, the limitation that a traditional wind field inversion method depends on a large-range overall consistency hypothesis is broken through, a local linear modeling method is adopted, small-scale changes, local non-uniformity and turbulence disturbance existing in an actual wind field can be better adapted, and the adaptability and modeling precision of an inversion model to a complex wind field structure are improved; the anti-noise capability, the error control and the inversion convergence are remarkably improved; the invention further provides a system, equipment and a storage medium for implementing the method.

The invention provides an efficient data acquisition method and system for an HPLC (High Performance Liquid Chromatography) dual-mode communication module based on edge calculation, and relates to the technical field of HPLC dual-mode communication data acquisition. The method comprises the steps of collecting multi-dimensional resource data of an HPLC dual-mode communication module in real time, performing feature extraction by adopting a local linear embedding algorithm, constructing a resource state evaluation model based on an extreme learning machine, evaluating a resource load state of the communication module in real time, and establishing a multi-target optimization model. Solving by adopting a non-dominated sorting genetic algorithm III to obtain an acquisition strategy solution set, selecting an optimal acquisition strategy, converting the optimal acquisition strategy into a control instruction, and adjusting data acquisition parameters in combination with predictive control and a feedback correction mechanism. According to the method, the problems of resource waste, single index and the like of an existing scheme are effectively solved by performing adaptive denoising, feature extraction and resource state accurate evaluation on the multi-dimensional data and combining multi-objective optimization of data space-time association mining and acquisition efficiency, energy consumption and reliability.

This application discloses an artificial intelligence-based method, apparatus, device, and medium for 3D cell structure reconstruction, relating to the field of cell reconstruction. The method includes determining cross-modal fusion features using a multi-scale feature encoder based on point cloud data in the same coordinate system and pre-defined biological prior features, including local feature extraction, global feature extraction based on multi-head relative self-attention, and cross-modal feature fusion; performing dimensionality reduction using an improved local linear embedding manifold learning algorithm based on the cross-modal fusion features; determining topology-enhanced high-dimensional features using a graph convolutional network based on spatial proximity, feature similarity, and biological relevance, using point cloud data in the same coordinate system, pre-defined biological prior features, cross-modal fusion features, and low-dimensional features preserving topological relationships; and performing 3D reconstruction using a decoder based on the topology-enhanced high-dimensional features. This application can improve the accuracy and mechanistic interpretability of 3D cell models.

The invention relates to the technical field of power monitoring, and discloses a power module operation state monitoring method and system, and the method comprises the steps: carrying out the multi-frequency calibration of a high-frequency electrical signal of a power module, and obtaining a standardized multi-dimensional electrical signal; decomposing the standardized multi-dimensional electrical signal into a limited bandwidth modal component, and performing Hilbert transform to obtain a high-dimensional state feature vector; performing local linear embedding dimension reduction on the high-dimensional state feature vector to obtain a low-dimensional reference coordinate; acquiring a low-dimensional monitoring coordinate of the current high-dimensional state feature vector; constructing a decision diagram of the low-dimensional reference coordinates, and performing clustering center-to-center spacing analysis on boundary points in the decision diagram to obtain a dynamic adaptive threshold value; state deviation degree analysis is carried out on the power module, difference comparison is carried out on an analyzed state abnormity index and a dynamic self-adaptive threshold value, and a state abnormity early warning signal of the power module is generated; according to the invention, the efficiency of monitoring the running state of the power module can be improved.

The application discloses a non-rigid point cloud registration method and relates to the technical field of three-dimensional reconstruction. 2 The application further discloses a non-rigid point cloud registration method, which comprises the following steps: standardizing data points of a collected non-rigid point cloud and reference point cloud; calculating a local linear embedding weight matrix L of the data points; calculating a matrix M of the data points according to the matrix L; calculating a Gram matrix G of the data points; calculating a corresponding matrix P of the data points based on the reference point cloud; iteratively calculating a non-rigid transformation coefficient matrix W by using an L-M algorithm based on the matrix G, the matrix M and the matrix P; updating λ and σ in the non-rigid transformation coefficient matrix W according to the number of iterations; calculating the matrix P and the non-rigid transformation coefficient matrix W again; setting an iteration termination condition, that is, the number of iterations reaches a set value or a relative error of a target function value is less than a threshold value; and outputting a non-rigid transformation T based on the matrix G and the non-rigid transformation coefficient matrix W after the iteration is terminated.

The invention discloses a high-speed train traction system fault detection method based on local linear embedding, and belongs to the technical field of fault diagnosis. The method comprises the steps that firstly, traction system sensor data are collected and preprocessed; secondly, mapping high-dimensional data to a low-dimensional space by utilizing a local linear embedding algorithm, and reserving a local geometric structure of the data; thirdly, reconstructing original data based on the data matrix of the low-dimensional space, calculating residual signals, and extracting fault features; and finally, judging whether the high-speed train traction system breaks down or not by comparing the evaluation function with a threshold value. According to the method, the limitation of processing high-dimensional and nonlinear data by a traditional method is effectively solved, the accuracy of fault detection is improved, fault early warning of the high-speed train traction system can be realized, and the method has a relatively high engineering application value.

The invention discloses a peak correction and characteristic scale standardization method for liquid chromatogram data. The method mainly comprises the following steps: acquiring chromatographic data of an experimental sample to be detected and preprocessing the chromatographic data; carrying out local linear transformation on the abscissa (X-axis) of each peak by adopting a local expansion and contraction method, and accurately aligning the peak center to the standard retention time; the peak information of each sample is corrected through Gaussian fitting, so that the change of the peak information before and after alignment is within a threshold range; and finally, carrying out X-axis standardization processing, unifying the time interval and the characteristic quantity of the samples, and comparing the data of each sample under the same scale. The method is used for eliminating the inconsistency of the chromatographic data caused by different instruments and experiment conditions, successfully correcting the peak information in the data and unifying the characteristic scale; therefore, the chromatographic data can be imported into the machine learning model more simply and accurately for further analysis.

The invention provides a robust graph comparative learning method and system for traffic adversarial attacks of the Internet of Vehicles. The method comprises the following steps: according to a pre-screening neighbor result of provided data and a one-dimensional structure entropy, calculating to obtain an optimal neighbor number of the provided data; on the basis of the optimal neighbor number, local linear reconstruction is used, neighbor weights are obtained through calculation, and a robust optimization graph structure is constructed; and according to the robust optimization graph structure, a difficult case mining method based on a double-memory bank is adopted, and through class prototype alignment of low-rank decomposition, reinforcement of robust embedding representation of the model is realized. According to the robust graph comparative learning method for the traffic adversarial attack of the Internet of Vehicles, provided by the invention, through a high-quality graph structure and GCL negative sample construction, the model robustness and the feature discrimination degree are balanced while the adversarial attack is resisted, so that in a complex Internet of Vehicles intrusion detection environment, the robustness of the model is greatly improved. And the robustness and the safety of the system are obviously improved.

The application belongs to the technical field of circuit automation detection, and specifically provides a secondary alternating current loop insulation monitoring method and system based on multi-modal machine learning, comprising: processing high-frequency pulse current signals generated by insulation defect faults of a secondary loop through multi-element variational modal decomposition to obtain time-frequency domain features of fault signals under different modes; selecting feature parameters for fault diagnosis, using a local linear embedding algorithm to reduce dimensions of feature parameters of fault time domain reconstruction signals and fault frequency domain reconstruction signals, taking the reduced historical fault time-frequency feature vectors and real-time fault time-frequency feature vectors as a training set and a test set respectively, and classifying through a support vector machine algorithm to obtain fault types and fault rates of each fault type after classification. The method and system utilize that insulation defects of a secondary loop will cause partial discharge, and the partial discharge will generate high-frequency pulse current signals, and analyze the insulation state of the secondary loop by extracting features of fault signals.

The application discloses a method for intercepting a maneuvering target based on linear quadratic differential game proportional guidance, which comprises the following steps: obtaining a meeting state quantity and constructing a locally linearized relative lateral dynamic equation rate based on the two-dimensional relative motion relationship between a pursuit side and a target; describing the locally linearized relative lateral dynamic equation as a finite time domain zero-sum linear quadratic differential game, constructing a lateral deviation and lateral velocity approximate state vector according to the meeting state quantity, and solving an optimal lateral control quantity of the pursuit side; corresponding the optimal lateral control quantity of the pursuit side with a proportional guidance law, calculating an instruction angular velocity of the pursuit side, and realizing online interception of the target. The application can realize real-time adjustment of the heading angular velocity of a fixed-wing unmanned aerial vehicle, and improve the success rate of online interception of a target rotor-wing unmanned aerial vehicle. The application can be widely applied to the technical field of maneuvering target interception.

The application discloses a kind of based on ST-GAN's multi-source data dynamic risk early warning method and system, the method includes the following steps: S1. constructing the spatiotemporal generation confrontation network suitable for spatiotemporal data characteristics, utilize the network to generate extreme precipitation data, realize spatiotemporal unbalanced data reduction, obtain the equalization data of final output;S2. the equalization data of final output with high-dimensional space variable is carried out multi-source data fusion, and nonlinear spatiotemporal information conversion equation, and by local linearization obtains linear approximation model, to predict future time series;S3. under the present situation that extreme rainfall data amount is relatively insufficient, neural network based on dual learning theory accurately learns the parameter of nonlinear spatiotemporal conversion, estimates extreme weather event.The application is through spatiotemporal generation confrontation network (ST-GAN) and multi-source data fusion engine, significantly improves the precision and efficiency of natural disaster warning.

The invention relates to the field of strain detection, in particular to a self-adaptive strain detection method and system and a medium. The method comprises the following steps: inputting grid point coordinates, noisy displacement data and a maximum smooth half window; calculating the overall noise level sigma through local linear regression analysis; extracting local strain curvature characteristic quantity through cubic polynomial regression; establishing a noise and curvature balance relation based on the sigma and the curvature characteristic quantity, and calculating an adaptive smooth half window; obtaining the strain estimation value of each grid point through linear polynomial regression, and finally obtaining the strain distribution of the whole field. According to the method, the optimal smooth window can be automatically adjusted without manual intervention, balance is achieved between noise suppression and detail reservation, the problem that a traditional fixed window method is insufficient in precision in a non-uniform deformation area is solved, high-precision full-field strain detection of a noise-containing displacement field is achieved, and the detection precision is improved. The method is suitable for derivative calculation of digital image related systems and various noisy continuous signals.

The present application relates to the field of strain detection, and particularly to a self-adaptive strain detection method, system and medium. The method comprises: inputting grid point coordinates, noisy displacement data and a maximum smoothing half window; calculating an overall noise level sigma through local linear regression analysis; extracting a local strain curvature characteristic quantity through cubic polynomial regression; establishing a noise and curvature balance relationship based on sigma and the curvature characteristic quantity, and calculating an adaptive smoothing half window; obtaining strain estimation values of each grid point through linear polynomial regression, and finally obtaining overall strain distribution. The method can automatically adjust the optimal smoothing window without manual intervention, balances between noise suppression and detail preservation, solves the problem of insufficient accuracy of traditional fixed window methods in non-uniform deformation areas, realizes high-precision full-field strain detection of noisy displacement fields, and is suitable for derivative calculation of digital image correlation systems and various noisy continuous signals.

The application discloses a linear variable parameter modeling method for time-delay industrial systems based on exponential optimal smoothing regularization, and belongs to the technical fields of system identification and industrial automation. The application aims at the problems that the modeling method adopted by the existing industrial systems estimates model parameters and time-delay parameters in steps, accumulates errors and causes low model precision. The application comprises the following steps: a local linear finite impulse response time-delay model is established, and a global model of the time-delay industrial system is obtained; a model identity hidden variable is introduced in a probability framework, a probability density function of global output of the global model is obtained based on local output distribution characteristics; a smoothing matrix is constructed based on an exponential optimal smoothing regularization method, and a prior distribution of local model parameters is obtained; an observation data set and a missing data set are established, an iterative updating formula of global model parameters is obtained based on a generalized expectation maximization algorithm, and global model parameter estimation values are obtained after the algorithm converges, so that the modeling of the global model is realized. The application is used for linear variable parameter modeling of time-delay industrial systems.

The invention provides a multi-scale time sequence feature extraction method based on a local linear layer. The method comprises the following steps: S1, carrying out standardization processing on an input univariate or multivariate time sequence; s2, setting a plurality of time windows with different sizes, wherein the size of each window corresponds to one scale; s3, under a single scale, retaining a local connection structure in a time dimension by adopting a mask mechanism, and independently extracting local features of time steps in each window in the scale through linear transformation; s4, under a single scale, sharing a linear weight in a channel dimension, and carrying out special extraction on all variables by adopting the same linear weight matrix; and S5, stacking the initial sequence and the feature sequence extracted under each scale along the dimension of the scale to form a multi-scale feature sequence of each variable, and outputting the multi-scale feature sequence to be used by a subsequent model. The invention provides a novel neural network layer named as a local linear layer, local modeling and multi-scale feature extraction can be realized, parameter efficiency and modeling capability can be improved, and the method is suitable for time sequence modeling tasks.

The invention relates to a dynamic measurement method for a convective heat transfer coefficient on the surface of heat exchange equipment, which specifically comprises the following steps of: (a) drawing up discrete points of the convective heat transfer coefficient, constructing a plurality of local linear heat transfer models and calculating a sensitivity coefficient; (b) establishing a temperature prediction model based on the model and the sensitivity coefficient, and obtaining a temperature prediction value of the measuring point; (c) comparing the predicted value with the measured value through rolling optimization to obtain a fluid temperature compensation quantity component; (d) constructing a weighting coefficient of the local linear heat transfer model according to the fluid temperature estimation deviation; (e) carrying out weighted integration on the discrete points, and outputting a convective heat transfer coefficient measurement value at the current moment; and (f) circularly updating the time domain until the measurement is finished. According to the method, the nonlinear system is segmented through local linearization, and rolling optimization and weighting strategies are combined, so that the surface convective heat transfer coefficient can be quickly and accurately measured.

The present application belongs to the technical field of image recognition, and particularly relates to a plate environmental protection coating surface defect identification method and system based on machine vision, which comprises the following steps: constructing a structure tensor matrix and calculating a local linearity factor; inversely modulating an original gradient vector by using the local linearity factor to construct an inverse modulation flow field; calculating the divergence of the inverse modulation flow field to generate a divergence potential energy diagram, and extracting a suspected defect center according to the local extreme value of the divergence potential energy diagram; and constructing a phase entropy index of the suspected defect center, and comparing the phase entropy index with a bubble determination threshold to determine a bubble defect. The present application can effectively suppress the interference of a high linearity wood grain background, distinguish bubble defects from surface stains and dirt, and improve the accuracy and robustness of plate coating surface defect detection.