1006 results about "Signal response" patented technology

The invention discloses an online course learning management method and system based on a knowledge graph, and the method comprises the steps: S1, analyzing course data, extracting knowledge points and an association relationship, and generating an initial knowledge graph signal containing knowledge point nodes and association edges; s2, collecting user interaction behavior data and extracting behavior features to generate behavior feature vector signals associated with knowledge point nodes; s3, detecting an abnormal learning mode, triggering a knowledge graph weight updating signal, and dynamically adjusting the associated edge weight; s4, updating the signal according to the initial knowledge graph signal and the knowledge graph weight; and S5, responding to the path optimization signal, and generating an interface sorting adjustment signal and a content push signal. According to the air quality intelligent monitoring system based on sensing data feedback, the problem that a traditional online course management system is single in data acquisition dimension and static and solidified in knowledge dependency relationship can be solved.

The invention relates to the technical field of grinding wheel monitoring, in particular to a centerless grinding machine full-life-cycle grinding wheel state monitoring system and method.According to the centerless grinding machine full-life-cycle grinding wheel state monitoring system and method, through combined difference extraction of a main shaft current mean value fluctuation ratio and a power gradient continuous variation value and assisted by transverse comparison of periodic response intervals, effective recognition of time period boundaries of working condition states is achieved; according to the method, operation stage division does not depend on static threshold judgment any more, a track trend is established through linkage judgment of an amplitude kick point and a slope extreme value, a dynamic evolution path of signal response is clearer, the identification resolution of abnormal changes and local deviation trends is improved, response synchronous identification of parameter state changes is achieved, and the identification accuracy is improved. In multi-signal collaborative analysis, directivity comparison between a temperature rise mutation fragment and a current residual error response is carried out, so that the consistency analysis capability of response logic between different monitoring paths is enhanced, and the real-time performance, stability and anti-interference capability of state identification are improved.

A wearable device configured to monitor physiological parameters of a wearer can include: a physiological parameter measurement sensor configured to monitor a plurality of physiological parameters; a hardware processor; a display in communication with the processor; and a band configured to secure the physiological parameter measurement sensor on a wrist of the wearer. In some implementations, the hardware processor is configured to: obtain a first plurality of signals from the physiological parameter measurement sensor when the band is secured on the wrist at a first tightness; determine a signal quality responsive to the first plurality of signals; and output an indication on the display to adjust tightness of the band with respect to the wrist from the first tightness to a second tightness based on the determined signal quality.

The invention discloses a relay state prediction and fault early warning method and system based on deep learning, and the method comprises the steps: S1, building a constraint condition of a generative adversarial network based on a relay physical model, and forming an enhanced fault waveform signal according with a physical rule through adversarial training; s2, receiving a real-time current and voltage signal and a mechanical vibration signal, and extracting an electric signal feature vector by using a time sequence convolutional network; s3, inputting the combined feature tensor into the lightweight assessment model, and outputting a health degree scoring signal; s4, responding to the meta-learning activation instruction, loading historical data of equipment to construct a parameter optimization set, performing online fine tuning on the early warning model based on a meta-learning framework, and generating a fault determination parameter; and S5, analyzing real-time signal characteristics according to the fine-tuned judgment parameters, and outputting graded early warning signals to a monitoring terminal. According to the method, the problems of early state prediction and accurate early warning of the relay under small sample fault data can be solved.

The invention relates to the technical field of water quality pollution detection, in particular to a water quality pollution rapid detection method driven by a quantum sensing technology. A quantum sensor array is deployed in a water area to be detected, and quantum signal response data of the water area is collected; constructing a pollutant quantum recognition model, and inputting quantum signal response data into the model to obtain an initial predicted value of pollutant concentration; generating an initial detection parameter of the quantum detection instrument based on the initial predicted value, executing primary pollution detection by the instrument according to the parameter, and collecting primary deviation data of an actual detection value and an expected detection value; determining a detection error according to the deviation data, further correcting the initial detection parameter, and generating an optimized detection parameter; and finally, the instrument performs secondary pollution detection based on the optimized detection parameters. According to the method, complex water sample pretreatment is not needed, water quality pollution detection can be rapidly and accurately completed, detection requirements of different water areas are met, and the detection efficiency and accuracy are improved.

The invention discloses an electrical stimulation type nerve regeneration method based on calcium signal regulation and control, and relates to the technical field of electrical stimulation type nerve regeneration, and the method comprises the following steps: based on a Schwann cell population calcium signal response feature set, identifying whether the calcium ion concentration is in a microliter stage which breaks through a lowest activation threshold and does not reach a stable state; judging whether the target nervous tissue is in a calcium signal activation critical state or not based on the behavior mode change of the Schwann cell population; based on the judgment result of the calcium signal activation critical state, an electrical stimulation adjustment control instruction is generated in real time in combination with the calcium ion concentration fluctuation trend and Schwann cell population response characteristics, and the calcium ion concentration stability is optimized and the activation state is maintained by adjusting the electrical stimulation intensity, the electrical stimulation pulse width, the electrical stimulation waveform type and the electrical stimulation application time sequence. According to the method, the problem that a calcium signal is difficult to recognize and dynamically optimize in a microliter stage is solved, and precise maintenance of Schwann cell activation and improvement of nerve regeneration efficiency are realized.

The invention discloses a brain deep nerve stimulation method and system based on adaptive adjustment, and relates to the technical field of brain deep nerve regulation, and the method comprises the steps: collecting a local field potential signal of a brain deep target region of a target patient, and extracting a beta frequency band power spectrum density and a gamma frequency band phase synchronization index as neural activity characteristic parameters; determining an individual baseline value and a preset threshold value based on historical data, and outputting a stimulation adjustment trigger signal when the beta frequency band power spectral density exceeds the individual baseline value and the gamma frequency band phase synchronization index is lower than the preset threshold value; in response to the trigger signal, calculating an optimal stimulation parameter combination through a gradient descent optimization algorithm and executing nerve regulation; and monitoring the signal change after regulation and control, calculating a relative change rate and updating a threshold value. Through a two-parameter joint judgment mechanism and a threshold updating strategy, individualized adaptive adjustment of stimulation parameters is realized, and the stimulation accuracy and the treatment effect are improved.

The invention relates to an unmanned aerial vehicle inspection method and system based on an overhead transmission line, and the method comprises the steps: S1, generating an unmanned aerial vehicle three-dimensional space coordinate signal, and outputting a position correction signal based on a tower topology database; s2, in response to the position correction signal, controlling a multispectral sensing module to synchronously acquire a visible light image signal, an infrared temperature signal and a laser radar point cloud signal of the line; s3, inputting the visible light image signal into a defect identification model to generate a first defect feature signal; s4, fusing the first defect characteristic signal and the second defect characteristic signal to form a cross-modal diagnosis signal, and triggering a defect classification decision signal; and S5, classifying decision signals and cross-modal diagnosis signals according to defects. According to the air quality intelligent monitoring method and system based on sensing data feedback, the problems of inaccuracy in routing inspection positioning, defect missing detection and insufficient endurance in a complex environment can be solved.

The invention discloses a cross-channel marketing automation method and system based on real-time behavior triggering. The method comprises the following steps: capturing a user behavior event flow in real time through a front-end acquisition unit; performing real-time mode matching on the event stream by utilizing a complex event processing engine, and generating a marketing trigger signal when a preset rule is matched; dynamically selecting an optimal channel for the user based on a real-time scoring model in response to the signal; the marketing content is sent to the selected channel through the unified execution module; and finally, tracking user feedback and updating model parameters and user preferences in real time to form closed-loop optimization. The corresponding system comprises a front-end data acquisition module, a streaming event processing module, a real-time intelligent decision-making module, a multi-channel execution module and a closed-loop optimization module which are used for implementing the steps. According to the method, millisecond-level intention perception, cost and experience optimal channel intelligent decision and data-driven adaptive optimization closed loop are realized, and the problems of delayed response, extensive channel and lack of self-adaption of traditional marketing are solved.

The invention relates to the technical field of debugging systems, in particular to an adaptive debugging system and method for touch screen FW parameters, and the system comprises a signal acquisition module, a trend triggering module, a temperature compensation calculation module, a window generation module and a parameter output module. According to the invention, the edge of the driving voltage is carefully collected, the sliding difference algorithm is introduced to filter out high-frequency noise, meanwhile, real-time sampling of environment temperature and baseline capacitance adjustment are combined, the influence of external variables on signals can be dynamically corrected, and accurate judgment of capacitance fluctuation is improved by utilizing periodic delay deviation and regression trend analysis. By means of main frequency feature extraction and zero crossing point sudden change detection, threshold setting can be refined, self-adaptive window expansion can be achieved, touch signal response is kept sensitive and consistent all the time, adjustment and optimization can still be automatically completed even under the condition that the temperature changes frequently or signal interference is complex, parameter drift and debugging failure are effectively prevented, and user experience is improved. And the parameter automatic matching efficiency and the finished product consistency in batch production are improved.

A method for validating operation of a driver integrated circuit includes providing a signal using an output node. The signal is provided using multiple set points in response to a change in state of an input signal. Each set point corresponds to a different phase of a multi-phase transition of the signal. The method includes providing a timer value at an end of a phase of the multi-phase transition and determining whether the signal is in a target signal range of the phase based on the timer value at the end of the phase, a predetermined value defining the target signal range of the phase, and a predetermined time limit for the phase. A current through the output node may be provided using the multiple set points, and a voltage on the output node may have the multi-phase transition.

The invention relates to the technical field of organic material detection, in particular to an imine molecular sensor with a five-detection function, two-component gel as well as a preparation method and application of the imine molecular sensor. The imine molecular sensor has three signal responses of fluorescence on-off, ultraviolet absorption and color change to Fe < 3 + >, Co < 2 + > and Ni < 2 + > in a solution state, and has ultraviolet absorption and fluorescence on-off dual-detection performance to acid and alkali changes; the imine molecular sensor and cholesteryl half ester can form two-component gel through a heating-low-temperature ultrasonic cooling method; after Fe < 3 + >, Co < 2 + >, Ni < 2 + >, NaOH and HCl are respectively added, the orange of the gel becomes milky white, reddish brown, transparent red, milky white and milky white under sunlight, and the fluorescence of the orange of the gel disappears under an ultraviolet lamp, so that the rapid and sensitive detection of Fe < 3 + >, Co < 2 + >, Ni < 2 + >, acid and alkali under two states of the imine molecular solution and the gel is realized. The imine molecular sensor provided by the invention has the advantages of mild synthesis conditions, simple preparation and the like, and creates favorable conditions for popularization and application of the imine molecular sensor.

The invention relates to the technical field of intelligent battery management, and discloses an intelligent battery hybrid parallel access system and a control method. According to the method, a battery pack is divided into a basic energy storage unit and a dynamic adjustment unit, a hybrid parallel access network is constructed based on a layered interaction architecture, unit voltage, loop current and temperature field distribution data are acquired through a multi-dimensional sensing module, feature extraction and weight fusion are performed by using a time sequence association processor, and a temperature field is obtained. And the input state decision module adjusts an access mode, and the two-way instruction analyzer executes mode switching. Meanwhile, real-time data and reference parameters are compared through a grading evaluation mechanism, after operation deviation is determined, an adjusting signal is transmitted through a feedback channel, an operation record is generated, and the response speed of the adjusting signal is in positive correlation with the deviation. The method comprises a hybrid architecture construction module, a multi-source data processing module and an operation deviation feedback module. According to the invention, efficient management and accurate control of the battery pack are realized, and the flexibility, stability and safety of the system are improved.

The invention discloses a fluorescent colorimetric sensor for detecting circulating tumor cells and a preparation method of the fluorescent colorimetric sensor. The preparation method comprises the following steps: S4, adding a prepared detection antibody modified with horse radish peroxidase and used for detecting the circulating tumor cells, and incubating; s5, adding a Tris buffer solution and a H2O2 solution, and incubating; s6, preparing an EuMOF nanoflower solution; and S7, adding the EuMOF nanoflower solution and 3, 3 ', 5, 5'-tetramethyl benzidine to react, detecting the fluorescence intensity in the presence of circulating tumor cells with different concentrations by using a luminoscope, detecting the ultraviolet absorbance value in the presence of circulating tumor cells with different concentrations by using a multifunctional microplate reader, and constructing a fluorescence intensity-concentration working curve and an ultraviolet absorbance value-concentration working curve. The method is used for qualitative and quantitative analysis and detection of the circulating tumor cells, realizes bimodal signal response to the circulating tumor cells, and has high sensitivity and simplicity and convenience in operation.

The invention discloses a power construction progress intelligent management method and system, and the method comprises the steps: carrying out the periodic cruise according to a preset height measurement position through an unmanned plane, and collecting a video stream of iron tower construction in real time; identifying the current construction height based on the video stream, comparing the current construction height with a preset height threshold value, judging that a construction progress node is reached when any threshold value is reached, and generating a trigger signal; responding to the signals to obtain coordinates of corresponding installation monitoring points, and controlling the unmanned aerial vehicle to collect videos at fixed points; component installation state analysis is carried out on the fixed-point video, and an abnormal alarm is generated when the feature conformity is lower than a threshold value; marking the timestamps of the video clips corresponding to the thresholds, and associatively storing the timestamps to a construction progress node; and generating a visual progress interface based on the association result. Through unmanned aerial vehicle cruise acquisition, video analysis and data association technologies, quantitative management of the construction progress, accurate detection of the installation quality, full-dimension tracing of the process and safety risk prevention and control are realized, and the construction efficiency and safety of the electric power iron tower are remarkably improved.

A tool for movement sequence analysis utilizing printed circuits. The tool determines a workflow sequence based, at least in part, on one or more printed circuits within a printed organic component. The tool activates the workflow sequence utilizing applied external stimuli on the one or more printed circuits. Responsive to activating the workflow sequence, the tool analyzes a signal response from the one or more printed circuits. The tool generates one or more printing recommendations based, at least in part, on the signal response analysis.

The invention relates to the field of display, and discloses a short-circuit protection circuit, an array substrate, a display device and a short-circuit protection method.The short-circuit protection circuit is used for an OLED device and comprises the steps that an information sensing module obtains the actual working current of the OLED device and real-time parameters related to the working state of the OLED device; the dynamic expectation calculation module calculates the dynamic expectation working current n of the OLED device according to the real-time parameters and a preset OLED electrical characteristic model; the decision module synthesizes the actual current, the expected current and the current change rate / fluctuation amplitude to generate a risk judgment signal; and the response control module executes protection operation according to the judgment signal. According to the OLED short-circuit protection method, the accuracy, the sensitivity and the intelligent management level of OLED short-circuit protection are remarkably improved through the technology of dynamically calculating the expected current and carrying out decision making and parameter self-adaptive adjustment by fusing the actual current time sequence characteristics.

The invention relates to the technical field of electronic connector performance monitoring, in particular to a connector performance detection method, and aims to solve the problem that a method for realizing performance early warning through dynamic correlation analysis of vibration signals and contact resistance in the prior art is not suitable for an L-shaped pin connector without a sheath. According to the method, micro-strain signals and contact resistance signals are synchronously collected, analysis time periods are divided in combination with plugging events, the plastic deformation accumulation degree and signal response delay characteristics are quantitatively evaluated, performance degradation coefficients are generated through fusion, and trend analysis is carried out; the problem that the relevance between the mechanical deformation and the electrical performance cannot be quantitatively evaluated in the prior art is effectively solved, and the method has the advantages of realizing early warning, improving the monitoring accuracy and reducing the maintenance cost.

The invention discloses a customer portrait analysis method and system based on a knowledge graph, and the method comprises the steps: S1, collecting a user behavior event flow and an external environment event flow, and generating a dynamic event signal carrying a timestamp based on a heterogeneous data source; s2, responding to a dynamic event signal, and analyzing an event type to trigger incremental updating operation; s3, in response to the real-time graph updating signal, updating the feature vector of the user node through an incremental graph calculation algorithm to generate a portrait updating signal; s4, generating a graph updating instruction and a portrait recalculation instruction to form a cooperative control signal; and S5, feeding back a graph updating instruction in the cooperative control signal to an incremental updating operation step, and feeding back a portrait recalculation instruction to a feature vector updating step to drive a real-time cooperative closed loop. According to the air quality intelligent monitoring method and system based on sensing data feedback, the problems of client portrait hysteresis quality and low map-portrait cooperation efficiency can be solved.

A wireless device may receive one or more radio resource control (RRC) messages comprising one or more configuration parameters. The one or more configuration parameters may comprise (i) a sounding reference signal (SRS) resource indicator (SRI), for a configured uplink grant, indicating an SRI-physical uplink shared channel (PUSCH) power control parameter set and (ii) a pathloss reference signal update parameter enabling an activation command to update pathloss reference signals of an uplink channel. In response to the one or more configuration parameters comprising the pathloss reference signal update parameter, the wireless device may determine a transmission power for a transport block of the configured uplink grant based on a pathloss reference signal mapped to the SRI-PUSCH power control parameter set. The wireless device may transmit the transport block with the transmission power.

Periodic in-field testing of system on chip functional units is described. In accordance with the described techniques, a scan pattern associated with a fault is applied to a functional unit during an idle event of the functional unit, the scan pattern defining a sequence of input signals. An output of the functional unit is received in response to the scan pattern. A status of the functional unit with respect to the fault is output based on the output of the functional unit and an expected output for the scan pattern.

The invention belongs to the technical field of oil and gas exploitation, and particularly relates to a low-yield well distributed optical fiber sound wave wellbore flow interpretation system and method based on thermal excitation. Downhole fluid is heated through a downhole heating device, and a thermal fluid slug with the temperature higher than that of unheated fluid is formed; the moving distance of the thermal fluid slug in the shaft in unit time is tracked by drawing an optical fiber strain response diagram, and the downhole fluid flow speed can be obtained. The shaft flow interpretation method is small in underground flow interference and obvious in optical fiber signal response, the technology is simple and easy to implement, and the problem that signals of the DAS in a low-yield well are not obvious can be effectively solved.

The invention discloses a fault signal dual-mode output method of a magnetic latching relay, particularly relates to the technical field of relay control, and is used for solving the problem of low mode switching efficiency caused by an existing physical jumper mechanism. The current mechanical state of the relay is detected by receiving a remote dual-mode switching instruction; applying a driving voltage sequence corresponding to the target mode according to the state; based on the iron core residual magnetism intensity and the dynamic margin of a target mode nominal value, judging and generating a driving enabling or degaussing trigger signal; when a driving enable signal is responded, the lag time difference between the voltage rising edge and the magnetic flux jump starting point is captured in real time, and the eddy current loss characterization quantity is generated by combining the product integration of the current differential and the temperature rise; adjusting the rising slope and the duty ratio of the driving voltage sequence according to the lag time difference and the over-limit condition of the eddy-current loss characterization quantity; and finally, the matching between an execution result and a target mode identifier is verified by controlling a coil current response waveform, and the continuous holding mode and the instantaneous pulse mode are remotely and reliably switched.

The invention relates to the technical field of infrared measurement, in particular to an intelligent calibration method and system for temperature measurement of an infrared sensor, and the method comprises the steps: enabling an infrared camera to align with a blackbody radiation source to form a calibration scene, collecting the output of a plurality of groups of pixels under different radiation temperatures and different internal part temperature states, fitting a temperature signal response model containing an exponential term and a bias term according to least square to obtain amplitude, exponential, temperature offset and output bias parameters; during measurement, a thermal image containing a target object and two reference blackbody sources is obtained, and parameters such as emissivity, environment equivalent radiation temperature, transmission medium and optical transmission, camera temperature and the like are synchronously obtained; performing non-uniformity and offset correction and abnormal pixel replacement on the thermal image; the initial temperature is inverted based on radiation transfer decomposition and compensation of internal radiation of the camera; and carrying out statistical averaging on the temperatures of the two blackbody reference regions to obtain a double-point correction coefficient, and outputting correction temperature or temperature distribution. According to the method, the problems that temperature inversion model parameters are difficult to adaptively update along with working conditions and on-site temperature measurement errors are difficult to effectively suppress when an existing infrared sensor array is in a long-distance non-contact temperature measurement scene can be solved.

The invention discloses a fast data protection method and device for SSD abnormal power failure, equipment and a medium, and relates to the field of SSDs, and the method comprises the steps: monitoring a power supply voltage, and generating an interrupt signal when the power supply voltage is lower than a preset threshold value; in response to the interrupt signal, turning off a power supply of a preset non-data path hardware module through the SSD controller; stopping scheduling all operation commands for the NAND flash memory through the SSD firmware, and emptying a data path in an NAND flash memory controller; writing the user data in the host write cache and the command information corresponding to the user data into a special physical address block reserved in the SSD in advance; after the SSD is powered on again, the address mapping table is recovered by scanning the write-in point in the NAND flash memory, the backup data and the command information in the special physical address block are read, and the backup data are written into the user data area according to the command information. According to the method, the key data backup time during abnormal power failure is shortened, the capacity requirement for the standby capacitor is reduced, and the problem that the cost is increased due to the fact that the backup time is too long is effectively solved.

The present invention relates to the technical fields of human-computer interaction and brain informatics, and in particular to a closed-loop neurofeedback method and device based on haptic memory electroencephalogram signal conditioning. The method comprises: collecting electroencephalogram signals of a subject, and performing real-time online data preprocessing; designing a haptic memory stimulation task, activating the haptic memory of the subject, and recording a corresponding electroencephalogram signal response; performing feature extraction and analysis on a corresponding electroencephalogram signal to obtain an electroencephalogram feature related to haptic memory; and on the basis of a haptic memory electroencephalogram feature signal, designing a closed-loop neurofeedback system, the closed-loop neurofeedback system monitoring the electroencephalogram signals of the subject in real time, performing real-time stimulation on the basis of a preset haptic stimulation task, and recording a corresponding electroencephalogram response. In the technical solution of the present invention, the haptic memory electroencephalogram feature signal is combined with the closed-loop neurofeedback system, thus providing new ideas and methods for the diagnosis and treatment of haptic memory-related diseases.

The invention relates to the technical field of artificial intelligence security detection, in particular to a detection method integrating deep feature extraction and attack recognition, which comprises the following steps: acquiring deep feature signal response, analyzing fluctuation and judging abnormity, recursively adjusting detection logic, updating snapshots, counting dependency sorting behavior priority, and judging detection timeout and adjusting priority. State reconstruction views are uniformly updated, and a cooperative detection mechanism is realized by identifying instructions. According to the method, by analyzing the fluctuation range of signal response in the deep feature space, detection deviation caused by abnormal signals can be recognized, classification of potential anomalies can be achieved, detection logic sequences of the anomalies and associated signals can be adjusted in a recursive mode, misassociation can be dynamically removed, a detection structure can be rebuilt, misjudgment interference is reduced, and the detection accuracy is improved. According to the method, key behavior processing of uncompleted detection is promoted, the detection rhythm and logic rationality are guaranteed, the detection process has dynamic coordination and multi-dimensional linkage capabilities, and the response efficiency and recognition accuracy of hidden attacks in a complex scene are effectively improved.

The invention relates to the field of biological medicine, in particular to a prodrug structure with ROS signal response performance and a preparation method and application thereof. The prodrug molecule with ROS signal response performance developed by the invention can quickly release living active molecules in a disease signal ROS-enriched environment, and plays a role in targeted therapy. The method has good universality, prodrug modification can be carried out on multiple molecules, four kinds of drug molecules approved to appear on the market are included, and it is verified that the solubility and biocompatibility of drugs can be improved, cytotoxicity can be reduced, and the effect of targeted therapy can be improved. Based on the structural characteristics of the molecules, the molecules can also be complexed with PVA hydrogel, and responsive controllable release of the medicine is further achieved.

The invention relates to the technical field of big data, in particular to a personalized recommendation method and device based on structure and behavior layering and a medium, and the method comprises the steps: extracting global features and graph structure information of interaction data of a user and an article; modeling a personalized portrait of the user, and generating a filter coefficient; carrying out hierarchical coefficient fusion, and executing mixed spectral domain atlas filtering to obtain a personalized atlas filtering score; obtaining a basic recommendation score through multi-source score fusion, and executing dynamic score modulation to obtain a final recommendation score; and generating a recommendation list based on the final recommendation score, and returning the recommendation list to the user. According to the method, through layered and two-dimensional personalized modeling, a recommendation engine can dynamically adjust the signal response of each user in a low frequency-high frequency and global-personalized spectral domain, so that the diversity, novelty and interpretability of a recommendation result are remarkably improved while the recommendation precision is ensured.

The invention relates to the technical field of defect detection of image recognition, and particularly discloses a mobile phone card holder processing defect detection method and system based on image recognition, and the method comprises the steps: collecting multiple frames of time sequence images, and carrying out the sub-pixel-level space-time registration, thereby eliminating the influence of mechanical vibration and displacement deviation; calculating an optical flow field, extracting a local strain tensor, an optical flow vorticity and a motion consistency entropy, and constructing a time-varying feature matrix; carrying out frequency domain decomposition and motion amplification on the characteristic matrix, and enhancing micro defect signal response; the dynamic time domain features and the static image features are fused, and a defect detection model based on deep learning is established; and finally generating a defect probability cloud picture, and outputting defect types and positions in combination with a dynamic threshold mechanism, so that high-precision and high-efficiency automatic detection of the small defects of the mobile phone card tray is realized, and the method has good stability and adaptability.