6201 results about "Test sample" patented technology

A multi-model fused avionic product health assessment method includes the following steps: collecting relevant data of an avionic product; performing data pre-processing on the relevant data to obtain first data and second data; training a plurality of base models on the basis of the first data; performing quantitative measurement and fusion on the plurality of base models to obtain an integrated model; and inputting into the integrated model the second data which serves as a test sample to obtain a health assessment result of the avionic product. A plurality of base models are integrated by using an AdaBoost algorithm, and a reference can be provided for a method based on data driving in terms of application in the health assessment, prediction and management of an avionic product.

The present invention relates to the technical field of coal mine water inrush disaster prevention and control, and relates to an artificial intelligence-based method for identifying a water inrush point in a mine and parameters of a simulation model, comprising the following steps: step 1: establishing a numerical model on the basis of observation data, and determining prior information of parameters to be identified including position coordinates of a water inrush point; step 2: on the basis of the numerical model and the prior information of the parameters, generating a training sample dataset and a test sample dataset for an alternative model; step 3: constructing and training an alternative model neural network; step 4: testing the accuracy of the alternative model; and step 5: executing a simulated annealing algorithm to identify the position of the water inrush point and parameters of a simulation model. By using the method provided by the present invention, the problem in accurate positioning of water inrush points in underground coal mines can be solved, which is of great significance in carrying out scientific disaster management and rescue efforts in a timely manner.

The invention relates to the technical field of image recognition, and discloses a flame-retardant material surface defect image recognition method, device and equipment and a storage medium, and the method comprises the steps: carrying out the multi-angle image collection and preprocessing of a flame-retardant material combustion test sample, and obtaining a standardized multi-view image data set; performing multi-model feature extraction and feature matching processing to obtain a material defect representation vector set and a defect semantic feature set; constructing a self-adaptive connection structure and a hierarchical defect map; performing multi-layer information transmission and topological relation explicit modeling through a depth map neural network to obtain a defect node depth representation set and a material defect relation matrix; domain invariant feature extraction and structural consistency constraint are carried out, material-independent defect type distribution and defect severity quantification results are obtained, defect features of all angles of the surface of the flame-retardant material can be comprehensively captured, remote interaction characteristics between defects are extracted, and the discrimination capability of defect feature representation is enhanced.

The invention belongs to the field of deep sea corrosion experiment equipment, and particularly relates to a multivariable coupling control deep sea corrosion environment simulation and monitoring device, which comprises an environment control unit, a deep sea corrosion environment simulation unit, a deep sea corrosion environment simulation unit, a deep sea corrosion environment simulation unit and a deep sea corrosion environment monitoring unit, and is characterized in that the environment control unit comprises a seawater environment cabin, a test sample is vertically arranged in the seawater environment cabin, and two ends of the test sample penetrate through and extend to the outer side of the seawater environment cabin; the pressure control system and the temperature control system are connected with the seawater environment cabin through sealing flanges and are respectively used for simulating deep sea hydrostatic pressure and a low-temperature environment; the mechanical loading system is used for bearing the seawater environment cabin and clamping the two ends of the test sample through the sample clamping system so as to apply a multi-axis dynamic mechanical load; the gas regulation and control unit is used for controlling the concentration of dissolved oxygen in the seawater environment cabin; and the multi-source monitoring module is used for collecting detection data and analyzing the detection data to obtain the corrosion rate of the test sample. The single-variable control limitation of a traditional device is broken through, and a high-fidelity experiment platform is provided for deep sea material corrosion mechanism research and equipment life evaluation.

The invention discloses a multi-modal ship target individual identification method and system, and relates to the technical field of ship target identification, and the method comprises the steps: obtaining multi-modal data of a target region, and constructing a training set and a test set; each training sample in the training set is input into a target individual prediction model, a multi-granularity adaptive loss function is adopted as a loss function to carry out model training, and the target individual prediction model adopts an image feature extraction method based on cross-modal guide enhancement to carry out feature extraction and splicing fusion on each training sample; and a target individual identification result of each training sample is obtained by adopting a visual angle self-adaptive fusion method of embedded type label information. And after training is completed, a trained target individual prediction model is obtained, each test sample in the test set is input into the trained target individual prediction model, and a final target individual identification result is obtained. According to the invention, the generalization ability of model learning individual features can be improved, and the accuracy of ship target individual identification is improved.

The invention relates to a deep learning model quantitative evaluation method and device and a storage medium, and the method comprises the steps: automatically determining a target quantification strategy through a preset mapping relation between a quantification strategy and model parameter information, does not need to manually debug parameters, carries out the model quantification of an optimized model based on a quantification tool chain configured by a configuration file, and improves the efficiency. According to the method, a quantitative model and a test sample are configured to obtain a deployment package, the deployment package comprises the quantitative model and the test sample, a reasoning index can be directly obtained on end-side equipment and compared with an original model index, objective quantitative evaluation is formed, and the method forms a standardized closed-loop process from model structure optimization, strategy generation and tool chain configuration to end-side deployment and evaluation. Manual intervention of connection of all links is reduced, dependence of an existing method on artificial experience is broken through, model quantitative evaluation can be automatically achieved, and the model quantitative requirement of a deep learning model is met.

A multispectral combined method for online detection of slurry components in a mineral processing process, relating to the technical field of spectrum detection. The multispectral combined method comprises: collecting slurry samples having a gradient distribution of moisture content to serve as test samples, and using an NIR system to acquire near-infrared spectrum data of the test samples; performing spectral preprocessing on the near-infrared spectrum data, so as to reduce the interference of spectral noise on subsequent modeling, and establishing an NIR regression model for slurry moisture content; using an LIBS system to acquire plasma spectrum data of the test samples, and establishing an LIBS regression model for slurry concentration; substituting the NIR regression model for slurry moisture content into the LIBS regression model for slurry concentration, and optimizing an algorithm to realize model correction and compensation; and establishing a quantitative analysis model for slurry element content to realize LIBS and NIR multispectral combination. The multispectral combined method can effectively improve the online detection precision of slurry components in a mineral processing process and improve the accuracy of a regression model.

The present disclosure provides an artificial intelligence-based method for enhancing mine safety by identifying and predicting locations of water inrush points in a mine, including the following steps: S1: constructing a numerical model to determine priori information of parameters to be recognized based on observation data, including coordinates of locations of water inrush points; S2: generating a training sample dataset and a test sample dataset of an alternative model based on the numerical model and the priori information of the parameters; S3: constructing and training a neural network of the alternative model; S4: testing an accuracy of the alternative model; and S5: performing a simulated annealing algorithm to identify the locations of water inrush points and simulation model parameters.

The variable excitation intensity time-resolved photoluminescence carrier lifetime test method comprises the following steps: S1, pulse excitation time-resolved photoluminescence response is carried out to obtain a time-dependent attenuation curve of the unbalanced carrier concentration of a to-be-tested sample, and S2, according to the attenuation curve, the time-dependent attenuation curve of the unbalanced carrier concentration of the to-be-tested sample is calculated. The carrier lifetime data under different excitation intensities are fitted to obtain the SRH lifetime and the surface composite lifetime of the sample, and the # imgabs0 # SRH composite lifetime, the # imgabs1 # radiation composite lifetime and the # imgabs2 # auger composite lifetime are obtained. The variable excitation intensity time-resolved photoluminescence carrier lifetime test method can accurately measure each component of the carrier lifetime.

The invention discloses a multi-source data-combined pit-pond water depth calculation and storage capacity calculation method, which comprises the following steps of: acquiring actually measured water depth data, performing data preprocessing, and dividing the data into a training sample point set and a test sample point set; acquiring a multispectral remote sensing image of the reckoning area, and executing image preprocessing; unifying to the same coordinate system; fractal dimensions of each wave band and the combined wave band are calculated, and a water body region is segmented; establishing a mapping relation between the water depth values of the training sample points and the standardized multispectral apparent reflectance; establishing a water depth inversion model, and verifying the water depth inversion model by using the test sample point set; calculating discrete water depth points of the pit pond by using the optimal water depth inversion model; and establishing a minimum energy curved surface to obtain the water storage amount of a single pit pond, and accumulating to generate the total water storage amount. According to the method, actual measurement and remote sensing data are fused, the energy minimization water depth inversion model is constructed, accurate calculation of the water storage capacity of the pit pond is achieved, and the method has the advantages of being high in precision and high in applicability.

The invention discloses an attention deficit hyperactivity disorder recognition system and recognition method, and belongs to the technical field of attention deficit hyperactivity disorder. The data processing module is used for carrying out preprocessing and feature extraction on the acquired electroencephalogram data; a multi-source feature fusion mechanism is firstly used for the extracted original feature data, and then a data enhancement strategy is applied; a multi-source fusion feedback regulation network model is constructed, wherein the model is of a CNN-GRU parallel modeling structure; the training module is used for inputting the enhanced data into a model for training, key hyper-parameters are dynamically adjusted by a performance feedback adjusting mechanism in the training process, and the performance feedback adjusting mechanism is used for dynamically adjusting key training parameters according to the performance of the verification set; and the classification module is used for classifying to-be-detected samples through the trained multi-source fusion feedback regulation network model and outputting a final recognition result. The ADHD electroencephalogram recognition method effectively improves the accuracy, robustness and generalization performance of ADHD electroencephalogram recognition.

The invention provides a double-space fatigue electron microscope test system, which is applied to the technical field of material analysis, and comprises an electron microscope module, a fatigue loading module and a vacuum isolation damping module, the vacuum isolation damping module controls two spaces of the electron microscope module and the fatigue loading module to be communicated or independent; the fatigue loading module comprises a hydraulic loading actuator, an electromechanical loading actuator, a loading rod and a fixing clamp, the fixing clamp is used for clamping a to-be-tested sample, and the loading rod is used for stretching or extruding the to-be-tested sample; and the fatigue loading module is driven by the moving mechanism to reciprocate along the X axis to adjust the position of the to-be-tested sample or rotate along the X axis to realize switching of different stations. According to the double-space fatigue electron microscope test system, combined use or independent use of the electron microscope module and the fatigue loading module is realized, and adverse factors such as vibration and pulse generated by the fatigue loading module during electron microscope observation are improved; in-situ compensation is achieved by means of cooperation of loading force and displacement of the fatigue loading module, and the observation effect of the electron microscope is improved.

The invention relates to the technical field of composite material detection, in particular to a fiber-reinforced thermosetting thermoplastic composite structure interface quality detection method and system.The method comprises the steps that vibration excitation is applied to a target area of the surface of a to-be-detected sample, and the amplitude of the excitation is lower than the damage critical value of a weak bonding interface; a dynamic response signal perpendicular to the interface is collected in a vibration field non-diffusion area of the excitation point of the target area; performing frequency domain transformation on the dynamic response signal, and separating a fundamental frequency component and at least one high-order harmonic frequency component; according to the nonlinear enhancement effect of the high-order harmonic frequency component relative to the fundamental wave frequency component, whether the weak bonding defect exists in the target area or not is judged. According to the method, the technical bottleneck that the dynamic evolution process of the weak bonding defect of the composite material interface is difficult to capture in real time is effectively solved, sensitive response to the microdefect is realized through a nonlinear harmonic characteristic enhancement effect, and the recognition precision of traditional ultrasonic detection is improved.

The invention relates to a rolling bearing fault diagnosis method and device based on a composite multi-scale attention entropy and an optimized SVM, and a medium, and the method comprises the steps: collecting vibration signals of a rolling bearing in different states, obtaining the damage size grade in each state according to the vibration signals, and obtaining a fault data set; performing coarse graining processing on the vibration signal to obtain a coarse grain sequence, and calculating an entropy sequence of the coarse grain sequence under different scale factors by adopting a composite multi-scale attention entropy algorithm improved by a fractional order algorithm; entropy values of a plurality of first scale factors in the entropy sequence are selected as bearing fault feature vectors; optimizing hyper-parameters of the support vector machine by using a collaborative group optimization algorithm; and constructing a support vector machine classification model based on the optimal hyper-parameter, and performing fault prediction and judgment on the test sample. Compared with the prior art, the method has the advantages of high noise robustness, high global convergence, high adaptability and the like.

According to the displacement modulation and optical assistance-based SERF atom magnetometer magnetic imaging method, an SERF atom magnetometer integrated with a magnetic flux guider is used as a magnetic sensor, and low-frequency magnetic field modulation is introduced to the magnetic sensor by controlling a to-be-tested sample to perform periodic infinitesimal displacement in the sensitive direction of the magnetic sensor; the modulation signal is demodulated by combining a phase-locked amplification technology, and a magnetic response signal subjected to phase-locked demodulation at each spatial position is extracted; constructing a magnetic response matrix through two-dimensional track scanning of micron-order step pitch; an image acquisition module and a reflection optical assembly are introduced for spatial registration so as to monitor the relative position in real time, and the visibility and controllability of the magnetic imaging process are guaranteed; the point-by-point detection of fT-level weak magnetic signals can be realized on the premise that an array structure is not needed; the low-frequency noise interference can be effectively inhibited, the stability and definition of direct-current and low-frequency magnetic field imaging are improved, and the method is suitable for high-resolution space imaging application of weak magnetic sources such as magnetic domain structures, nervous tissue slices and ancient rock samples.

The invention provides a cable bending fatigue life prediction method and system based on reinforcement learning, and relates to the technical field of cable life prediction, and the method comprises the steps: collecting cable test sample data, calculating a stress distribution field and a strain distribution field, constructing a stress intensity factor calculation module, and calculating a crack propagation track and rate. A damage evolution tensor is constructed as a state feature, a network is trained based on a deep Q learning algorithm, uncertainty is quantified by adopting a Markov chain Monte Carlo method, and a credible interval and an early warning threshold are established.

The invention discloses a low-illumination image enhancement method based on a multi-scale fusion U-Net network structure. The method comprises the following steps: S1, constructing a training data set and a test sample set; s2, constructing a low-illumination enhancement network model; s3, jointly calculating a plurality of loss functions; s4, training a low-illumination enhancement network model; s5, performing conversion quantitative deployment on the trained low-illumination enhanced network model; s6, preprocessing the image data to be enhanced; s7, inputting the preprocessed image data into the low-illumination enhancement network model for image enhancement; and S8, outputting an enhanced normal illumination image by the low illumination enhancement network model. According to the method, on the premise of ensuring high accuracy, the parameter quantity of the model is reduced, the portability of the model is enhanced, and the low-illumination image enhancement technology has a good development prospect in the embedded field.

The invention discloses a fault diagnosis method for an aviation hydraulic pump, belongs to the field of aviation hydraulic pumps, and aims to solve the problems of lack of feature information and low diagnosis precision caused by insufficient samples in an existing method. The method comprises the following steps: step 1, constructing a double-source feature extraction module, and respectively extracting time-frequency domain prior features and high-dimensional features of vibration signals; the extracted features are used for calculating a prior feature prototype and a high-dimensional feature prototype of a fault category; step 2, constructing a fault prototype strengthening module, and based on a self-attention fusion mechanism, performing weighted fusion on the prior feature prototype and the high-dimensional feature prototype to generate a fault strengthening prototype; and step 3, constructing a small sample fault classification module, measuring the distance between the test sample and the fault enhancement prototype through cosine similarity, outputting a fault category corresponding to the maximum similarity, and completing fault classification.

The invention discloses a composite shielding material effect experiment data analysis method, and particularly relates to the field of material shielding effect analysis, and the method comprises the following steps: measuring key data such as dielectric constant, magnetic conductivity, conductivity and thickness of each layer of a composite shielding material through an electromagnetic performance measurement instrument; establishing an electromagnetic wave propagation model for describing propagation of electromagnetic waves in the multilayer medium based on a Maxwell equation, and defining boundary conditions of a test sample; time domain and frequency domain characteristics of an electromagnetic interference source are simulated through Fourier transform, and actual input interference is simulated. And analyzing the dynamic response of the electromagnetic wave propagation model under transient and steady-state interference signals to form an electromagnetic wave interference response propagation model. Through comparison with actual shielding effectiveness test data, model parameters are adjusted, and it is ensured that model calculation errors are within an acceptable range. A full-band shielding effectiveness prediction map of the composite shielding material is generated based on the model, and theoretical support is provided for material design and optimization.

The invention provides an emotion recognition method and system based on evidence theory multi-modal data decision fusion. According to the method, two cluster centers of each piece of data are clustered through a PCM clustering algorithm, and the two cluster centers are used as an upper bound and a lower bound to construct an interval number model, so that the effect of the interval number model in a noise environment is improved. And a recognition framework is constructed, and the support degree of data features is calculated according to the interval length, so that the accuracy of the generated BPA during classification is improved. The BPA of each sample is calculated by using the similarity and the feature support degree of the test sample and the interval number model, and then evidence combination is carried out through a Dempster combination rule. Secondly, the conflict between the evidences is judged by using a Jousseme distance, then the similarity between the evidences is calculated to construct a similarity adjacent matrix, then the relationship strength between the evidences is defined by using the sum of the similarities, the weight of the conflict evidences is calculated and modified by integrating the direct relationship and the indirect relationship, and the original evidence body is subjected to weighted correction through the overall weight; and finally, combining through a Dempster synthesis formula to obtain a final result.

The invention provides a method for detecting silicon single crystal COP free area distribution, and relates to the technical field of silicon single crystal quality detection.The method comprises the steps that a silicon wafer to be detected is cleaned to obtain a first silicon wafer; uniformly smearing a transition metal ion solution on the surface of the first silicon wafer to obtain a second silicon wafer; placing the second silicon wafer in a high-temperature environment, and performing heat treatment under three temperature gradients to obtain a third silicon wafer; wherein in the first stage, transition metal ions are diffused to the surface layer of the silicon wafer to activate oxygen precipitates in a Pv region, in the second stage, the transition metal ions are promoted to be diffused into the silicon wafer to develop a Pi low-oxygen region, and in the third stage, nuclei coalescence in a B-band region is activated at high temperature to form a differential arrangement ring; cleaning the third silicon wafer through a mixed acid solution to obtain a fourth silicon wafer; and placing the fourth silicon wafer in a preferential corrosion liquid for preferential corrosion, and carrying out COP free area distribution detection on the obtained detection sample wafer. According to the scheme, the definition of each region boundary can be improved, so that the detection accuracy of COP free region distribution is improved.

According to the three-point bending high-temperature fatigue test instrument and method based on X-ray in-situ imaging, an extreme thermal field is applied to a sample based on halogen lamp infrared radiation focusing heating, a clamp ascends and descends through a servo hydraulic cylinder, and axial displacement of the ends of the sample is converted into bending moments at the two ends through the rotation characteristic of a connecting rod mechanism; therefore, a monotonous or cyclic bending external load is applied to the sample. Meanwhile, an oxidizing or inert test gas atmosphere is provided. And the acoustic emission system monitors the material damage failure starting time in real time. The device is small in size, high in precision and fast in response, and provides technical support for the damage evolution process characterization and failure damage mechanism research test of the tested sample under the heat-force-oxygen multi-field coupling action by combining a three-dimensional high-temporal-spatial-resolution X-ray characterization technology.

The invention relates to the technical field of chemical production process control, and particularly discloses a production process export quality prediction method based on performance-driven knowledge distillation, which comprises the following steps: step 1, acquiring observable process data of an industrial production process and test data of export quality indexes, and forming a data set; 2, cleaning and standardizing the data set, and randomly dividing the data set into a training set, a verification set and a test set according to a certain proportion; 3, building a basic quality predictor, and embedding an LSTM module into the basic quality predictor through a twinning neural network framework; 4, training a plurality of basic twin LSTM quality predictors to obtain a teacher model; step 5, training student models with the same structure by taking performance as drive and combining knowledge distillation and supervised learning; and step 6, inputting process data of the test sample, and obtaining an output predicted value, so as to improve the accuracy of the soft measurement of the production quality index in the industrial process on the basis of ensuring the stability of the model.

The invention discloses a continuous casting machine zero section base fault diagnosis method and system based on an improved Transform. The method comprises the steps that pressure signals including the normal state and the common fault state of a base are obtained; fusing the pressure signal features to construct enhanced multi-domain features; a nonlinear convergence factor, an adaptive weight and a Cauchy disturbance strategy are introduced to improve a whale optimization algorithm, and an IWOA model is obtained; a time sequence convolution fusion Transform network model is constructed, and IWOA is used to optimize parameters of a fusion convolution Transform, so that an optimal model can be obtained; and training the optimized fusion convolution Transform model, and inputting the multi-domain features of the test sample set into the trained network model to realize accurate identification of multiple states of the base. According to the method, the problem of model parameter selection is solved, efficient expression of features is realized, and the accuracy of base fault recognition is improved.

The invention discloses an aging test method, particularly relates to the field of aging tests, and solves the problem that an aging test result is inaccurate due to the fact that a current shutdown circuit breaker aging test method excessively depends on previous data, and the aging test method specifically comprises the following steps: selecting a fixed number of shutdown circuit breaker screen model machines as test samples; constructing a first test environment and a second test environment of the test sample; performing a hygrothermal aging test on the test sample based on the test parameters of the test environment to obtain a hygrothermal aging normal signal or a hygrothermal aging abnormal signal corresponding to the test sample; performing an ultraviolet aging test on the test sample based on the test parameters of the test environment to obtain an ultraviolet aging normal signal or an ultraviolet aging abnormal signal corresponding to the test sample; the aging test result of the test sample is judged based on the test passing rate. The aging test result of the shutdown circuit breaker screen model machine is accurately analyzed.

The invention provides an intrinsic solid-state ion diffusion coefficient test method and device based on optical imaging, and relates to the technical field of solid-state ion test.The method comprises the steps that a test sample is pretreated; carrying out a photoelectric test on the test sample; placing the electrochemical cell in an optical imaging system, connecting the electrochemical cell with an electrical testing system, applying a constant voltage deviating from a steady state to the monodisperse electrode plate, collecting the real-time change condition of the material under the modulation of the constant voltage in real time through an industrial camera, and outputting relational data of optical intensity and time; analyzing the diffusion path through in-situ characterization equipment, and obtaining an ion diffusion path corresponding to the to-be-detected ion battery electrode material particles in combination with dynamic information provided by an optical imaging system; and determining the characteristic time according to the relational data of the optical intensity and the time, and determining the intrinsic solid ion diffusion coefficient corresponding to the test sample by combining the characteristic time and the ion diffusion path corresponding to the to-be-tested ion battery electrode material particles.

The invention provides a sound wave generator soot blowing test method and system. The method comprises the steps that an experiment platform for simulating the working environment of a low-low-temperature economizer is built, and a test sample and sensing equipment are installed; setting power output parameters including a fixed power level and pulse power modulation parameter, a soot blowing frequency and phase control parameter, a time period and adaptive adjustment parameter, a sound wave frequency and an airflow velocity and vector control parameter; soot blowing operation is carried out, data are collected through sensing equipment, the soot coverage rate and particle information are obtained through an optical imaging sensor, and the pressure intensity generated by sound waves at the position where soot particles are located is obtained through an acoustic sensor array; calculating cleaning efficiency, and performing adjustment and iterative experiment on parameters according to the cleaning efficiency and data feedback; comprehensive optimization of the soot blowing process of the sound wave generator is achieved, the soot blowing effect of the low-low-temperature economizer is improved, then the overall operation efficiency of energy equipment is improved, the service life of the equipment is prolonged, and the requirements for environmental protection and energy saving are met.

The invention belongs to the technical field of biological simulation, and discloses an anus cleaning effect evaluation method and system.The method comprises the steps that an execution assembly is controlled to wipe an anus anatomical structure with a to-be-detected sample according to a target wiping strategy, and a multi-mode sensing assembly is controlled to collect target parameters on the anus anatomical structure; obtaining target monitoring data; according to the residual dirt amount in the target monitoring data, a cleaning efficiency score is obtained through calculation, a target mechanical simulation model is called, according to the humidity permeation gradient value and the pressure value in the target monitoring data and the to-be-detected sample, a skin stimulation risk score is obtained through calculation, and the skin stimulation risk score and the cleaning efficiency score form a cleaning effect evaluation report. Simulation and prediction of the anus cleaning process are achieved, the simulation test effect and accuracy are greatly improved, and a cleaning effect evaluation report is given to illustrate the safety of the to-be-tested sample.

The invention discloses a feature hijacking-based unified visual language model backdoor attack method. The method comprises the following steps of: obtaining a data set and a unified visual language model; initializing back door modules such as a multi-mode trigger, a trigger detector and a dynamic feature alignment module; constructing a harmful data set, randomly selecting a part of samples from the training set, injecting a multi-mode trigger into the samples, generating poisoning samples, and mixing original training data and generated poisoning sample data to generate the harmful data set; training the model by using a harmful data set, freezing original parameters of the model, and only allowing the backdoor module to participate in training; a poisoning model generated through model reasoning and training is normally expressed on a benign test sample, but when a text and an image backdoor trigger exist at the same time, the model outputs a preset answer, and backdoor attack is achieved. According to the method provided by the invention, the problem that the attack effect on the unified visual language model is insufficient due to single-mode triggering or insufficient feature disturbance is solved.

The utility model relates to the technical field of wave-absorbing material testing, and discloses a wave-absorbing material flat plate reflectivity testing device which comprises a flat plate sample table and a horn antenna matched with the flat plate sample table, the flat sample table comprises a cylindrical rod, a base, a rotating track, a table top and an inverted T-shaped metal panel which are connected with one another; the inverted T-shaped metal panel is driven to rotate along the circumferential direction when the rotating track on the table top rotates; the testing sample is arranged on one side face, facing the horn antenna, of a vertical panel of the inverted-T-shaped metal panel, and when the rotating track rotates in the circumferential direction, the inverted-T-shaped metal panel is driven to rotate in the circumferential direction, or the angle between the vertical panel and a horizontal panel of the T-shaped metal panel is adjusted, so that the testing sample is tested. And direction angles and pitch angles between the plane of the test sample and the horn antennas can be adjusted. The device has the advantages of being flexible and convenient to use, small in occupied area, capable of achieving multi-stage adjustment, high in testing precision and the like.