33 results about "Abnormal distribution" patented technology

The invention relates to a rubber part finished product forming degree detection method, system and equipment and a medium, and belongs to the technical field of rubber part forming degree detection. The molding degree detection method comprises the following steps: receiving surface point cloud data and surface geometric position information of a to-be-detected rubber part; constructing a three-dimensional surface contour model of the rubber part to be detected; mapping the surface pressure distribution data to the three-dimensional surface contour model, and detecting a pressure abnormal distribution area; extracting light scattering characteristic data according to the optical image data; comparing the light scattering characteristic data with the pressure abnormal distribution area to obtain a defect area analysis result; dynamic deformation characteristics of the rubber part to be tested under the dynamic environment simulation test are obtained; marking the three-dimensional surface contour model, and constructing a comprehensive feature model of the rubber part to be detected; and carrying out difference analysis on the comprehensive feature model to obtain a feature difference result of the to-be-detected rubber part, and generating a forming degree scoring report of the to-be-detected rubber part. According to the invention, the molding degree of the rubber part can be accurately detected.

The present invention relates to the technical field of intelligent supervision. Disclosed are a distribution-network-device state simulation method and system based on fault coupling analysis. The method comprises: collecting state data of a distribution network device, and pre-processing the collected data; performing analysis and mining on the state data of the distribution network device, and extracting a device fault mode, a coupling relationship and an anomaly; providing a visualized analysis result, and providing real-time monitoring and alarming functions, such that the anomaly of the distribution network device is discovered in a timely manner; and formulating an intelligent scheduling algorithm to search for the specific location of the abnormal distribution network device. The present invention can implement decision-making and planning of anomalies, thereby facilitating the timely discovery of anomalies in distribution networks.

The invention provides a physical examination data exception association analysis method and system based on a knowledge graph, and relates to the technical field of exception association analysis, and the method comprises the steps: determining all physical examination items of a target physical examination center; aiming at a plurality of selectable physical examination indexes, executing abnormal generality association and generality confidence degree analysis among the indexes, and establishing an abnormal generality association map; receiving a physical examination report; executing abnormal generality association verification based on the abnormal generality association graph; performing abnormal distribution characteristic analysis in a preset period for the plurality of selectable physical examination indexes, and performing distribution abnormality verification on the physical examination report; and performing joint abnormity verification according to the distribution abnormity verification result and the generality verification result, then performing abnormal physical examination item positioning, and sending the abnormal physical examination item positioning to a target physical examination center for reminding. The technical problem that potential health risks are difficult to recognize in time and the accuracy of physical examination results is affected in the prior art can be solved, and the technical effect of improving the accuracy of physical examination anomaly detection is achieved.

The invention relates to an atrial fibrillation radiofrequency ablation postoperative recurrence prediction method based on multi-modal analysis. Comprising the following steps: dynamically capturing the form and spatial distribution of a low-voltage region by using a three-dimensional electroanatomical map; introducing a regional complexity quantitative index, and analyzing the non-linear degree of the boundary of the low-voltage region; grading the voltage intensity gradient in the low-voltage region, and quantifying the sensitivity to postoperative recurrence; establishing a composite index model of the area ratio of the low-voltage region and the fragmentation potential intensity; based on spatial distribution and voltage features, performing high, medium and low recurrence risk classification on the low-voltage region, and outputting a classification heat map; analyzing the migration path of the fragmentation potential in the three-dimensional space in the operation, and extracting the migration speed, direction and frequency change; discovering a fragmentation potential concentration trend of the high recurrence sensitive area by using a trajectory clustering algorithm; and extracting spectrum characteristics of the fragmentation potential signal based on wavelet transform, and identifying abnormal distribution of the high-frequency fragmentation potential.

The present application relates to a thrust bearing full-life fault diagnosis method and system based on film thickness and wear. The method comprises: first establishing failure probability models for two stages, i.e., a lubrication failure stage and a wear failure stage, and simultaneously representing a lubrication state and a wear state of a bearing in the form of probabilities; then, performing a normal distribution test on monitoring data, and performing normal distribution conversion on abnormal distribution data; and finally, by using the established failure probability models for the two stages, i.e., the lubrication failure stage and the wear failure stage, and the acquired normal distribution data, performing diagnosis on the lubrication state and the wear state of the bearing on the basis of a 3σ criterion, thereby realizing full-life fault monitoring. In the present invention, failure evolution situations of two stages, i.e., a lubrication stage and a wear stage, of a thrust bearing are represented in the form of probabilities, so as to realize the full-life fault diagnosis of the thrust bearing, thereby addressing the current gap in research on dynamic fault diagnosis throughout the full-life cycle of thrust bearings, and improving the accuracy of fault early warning and condition-based maintenance of the thrust bearings under actual working conditions.

The invention relates to an operation risk monitoring method, device and equipment for a distributed service cluster. Comprising the following steps: acquiring multiple types of operation data of each service node in a distributed service cluster, grouping each service node according to each type of operation data, determining a trained abnormal distribution proportion corresponding to the operation data characteristic corresponding to each group, and determining the abnormal distribution proportion corresponding to the operation data characteristic corresponding to each group; calculating an index value corresponding to each group according to the anomaly distribution proportion, then summing the index values of the groups corresponding to the operation data of each type to obtain an anomaly score corresponding to the corresponding data of each type, and finally summing the anomaly scores of all types to obtain a total anomaly score. And predicting the operation risk of the distributed service cluster according to the total abnormal score, thereby realizing quantification of the abnormal condition of the distributed service cluster, and accurately predicting the operation risk even if the number of the computing nodes in the distributed service cluster changes dynamically.

The invention provides a value-added service abnormal distribution monitoring method and device, equipment and a medium, and the method comprises the steps: obtaining the request information of a target user for a target value-added service, and generating a target user portrait of the target user according to the request information; determining a first matching result of the portrait of the target user and a preset portrait library; when the first matching result indicates that the abnormal user portrait similar to the target user portrait does not exist in the preset portrait library, determining a service request attribute of the target user; determining a second matching result of the service request attribute and a preset early warning rule; and generating early warning information of abnormal distribution of the value-added service according to the second matching result. Therefore, automatic monitoring and early warning of abnormal distribution of the value-added service can be realized, so that the timeliness and accuracy of processing abnormal distribution of the value-added service are effectively improved.

The invention relates to the technical field of leprosy nerve injury grading evaluation, in particular to a leprosy nerve injury intelligent grading evaluation method based on high-frequency ultrasound multi-parameter fusion, which comprises the following steps: acquiring multiple frames of nerve images by adopting high-frequency ultrasound, analyzing a gray scale trend to construct a closed path, and screening out frames with abnormal distribution to generate a consistent sequence; and extracting a region coverage range, a brightness form and a texture arrangement feature, generating a change description according to a feature difference, matching a preset grading rule according to a description type, judging a damage level, and obtaining an intelligent grading evaluation result of the leprosy nerve injury. According to the method, the neural region is adaptively defined by using the image gray scale change, experience dependence is reduced, the consistency image sequence is screened, the spatial homology and stability of feature comparison are ensured, the single index interference is described and weakened based on the overall trend organization change, and the continuity and distinction degree of the evaluation result under the complex background are ensured. And the applicability of the grading conclusion in the dynamic observation scene is enhanced.

The invention provides an intelligent overload early warning method and system for a charging extension socket. The method comprises the following steps: synchronously sampling current and voltage of each socket, converting the current and voltage into digital signals, and obtaining an electricity consumption sequence of each socket; calculating the load proportion of each socket, and if the proportion exceeds a first preset threshold value, judging that abnormal distribution exists and extracting abnormal distribution characteristics; generating an abnormal oscillation feature set based on the abnormal distribution features; classifying the abnormal oscillation feature set, calculating an abnormal probability in combination with historical statistical data, and determining an abnormal oscillation state if the probability is higher than a second preset threshold value; and a decision tree algorithm is adopted to obtain a correlation intensity value of the abnormal oscillation state and the load proportion, incremental detection analysis is executed, and if an existing value in an incremental change sequence exceeds a third preset threshold value, an overload early warning signal is generated. According to the invention, through accurate load monitoring, efficient abnormal oscillation identification and dynamic correlation analysis, intelligent overload early warning is realized, and the safety of the charging extension socket is improved.

The invention discloses an intelligent measurement and control power distribution data processing method, system and device and a medium, and relates to the technical field of data processing. The method comprises the following steps: acquiring current network load information of a measurement and control power distribution network, and adjusting sampling frequencies of a plurality of data nodes in the measurement and control power distribution network based on the network load information to obtain a corresponding target sampling frequency; receiving first power distribution data collected by the data nodes under the adjusted sampling frequency, and extracting corresponding fluctuation characteristics from the first power distribution data; comparing each fluctuation feature with a preset data feature reference set, and determining a target data node with abnormal power distribution data and an abnormal type corresponding to the abnormal power distribution data; and concurrently correcting the abnormal power distribution data in the target data node based on the abnormal type to obtain second power distribution data. By implementing the technical scheme provided by the invention, the effect of improving the processing efficiency of the measurement and control power distribution data is achieved.

The present invention discloses an intelligent distribution management system and method for labeling blood collection tubes based on the Internet of Things, which relates to the technical field of blood collection management. The management method comprises the following steps: collecting data on the labeling process of each blood collection tube by an automatic blood collection system to generate a labeling record; conducting a multi-dimensional quality assessment on any labeling record; tracking the distribution of the blood collection tubes and identifying abnormalities in the distribution; conducting a secondary quality assessment on the blood collection tubes with abnormal distribution, analyzing abnormalities in the labeling process and the distribution process of the blood collection tubes; analyzing the impact of abnormalities between different processes; formulating a corresponding distribution strategy for any blood collection tubes according to the impact between different abnormalities; conducting a quality assessment on the real-time labeling records generated in real time, and predicting changes in the quality of the blood collection tubes; marking the blood collection tubes with abnormalities in the predicted results and sending a risk warning.

The application discloses a network abnormal data security early warning evaluation processing system, comprising a risk feature acquisition module, a first profiling module, a second profiling module and a distribution correlation module; the first profiling module can evaluate the abnormal risk of traffic through a network neural model to obtain a first risk profile of encrypted traffic; the second profiling module can obtain the frequency abnormality of distribution detection features through global frequency abnormal distribution and local frequency abnormal distribution of the distribution detection features; the distribution correlation module can obtain distribution correlation features associated with the distribution detection features of the frequency abnormality, evaluate the degree of frequency abnormality in combination with the deviation degree of the distribution correlation features, and obtain a second risk profile of encrypted traffic corresponding to the distribution detection features; and the first risk profile is adjusted according to the obtained second risk profile. The application increases the monitoring accuracy of traffic risk and improves network security.

The application discloses a kind of detection method and system of inductor, define the abnormal part of the inductor at different assembly stages based on the identification of detection image of different assembly stages;Based on a plurality of abnormal parts and the corresponding position of inductor, construct the abnormal set of the inductor, and construct the abnormal distribution map of inductor according to the abnormal set and the factory image of inductor after production is completed;Further, according to the abnormal distribution map of inductor and the production batch of inductor, determine the process to be optimized of the production batch;For the abnormal problem corresponding to the process to be optimized, and output the corresponding optimization strategy for abnormal problem;According to the process to be optimized and the corresponding optimization strategy, output the corresponding optimization report, define the optimization degree of abnormal part in inductor based on each optimization report and the abnormal distribution map of inductor, ensure the accuracy of the optimization degree of abnormal part.

The invention relates to the technical field of wagon operation monitoring, in particular to an intelligent detection method for an unbalance loading state in the operation process of a railway wagon, which comprises the following steps of: acquiring an instantaneous load response track of each side axle box relative to a track in the operation process to form a left-right path differential response sequence; the method is used for reflecting wheel-rail contact abnormal distribution characteristics caused by unbalance loading. Inputting the left and right path differential response sequence into an abnormal disturbance code construction module, marking key load disturbance points by using a multi-window trajectory disturbance detection mechanism, and generating a corresponding adaptive disturbance code sequence; and outputting the unbalance loading state type of the current vehicle and the corresponding offset side direction. According to the method, the consistency judgment of the disturbance form and the physical configuration is realized, the unbalance loading grade and direction two-dimensional diagnosis result is finally output, and compared with a traditional threshold value method, the method has the advantages of low false alarm rate, high adaptability and good interpretability, and is suitable for various railway wagon marshalling structures and operation conditions.

The present invention provides a method for identifying arterial waveform signal artifacts of blood pressure, which can quickly and accurately identify arterial waveform signal artifacts in blood arterial waveform signal, and the result is accurate and not affected by personal factors of the analyst. The technical solution of the present application determines the scattering degree Sp of the arterial peak interval of the arterial waveform to be analyzed by the Lorenz scatter plot, and the irregular distribution abnormality of the arterial waveform peak on the time axis is judged; by finding the effective area of ​​the arterial waveform to be analyzed, the arterial waveform data to be analyzed with abnormal peak points and valley points is judged as an interference signal; by the mirror data of the data of the effective area of ​​the arterial waveform to be analyzed, the peak amplitude disturbance rate PError of the arterial waveform to be analyzed is calculated and determined, and the proportion of the number of abnormal peak points in the arterial waveform to be analyzed is determined; finally, the arterial waveform interference data is determined by the two conditions of Sp and PError. At the same time, the present invention also discloses a method for evaluating the quality of blood pressure arterial waveform signal.

The invention discloses an alarm signal management and control method, device and equipment and a storage medium, and relates to the technical field of information processing, and the method comprises the steps: obtaining a to-be-sent alarm signal; determining an abnormality judgment condition according to the index type corresponding to the alarm signal to be sent; determining anomaly distribution and a historical baseline value according to the anomaly judgment condition, the index type and the generation time of the alarm signal to be sent; and performing signal control on the alarm signal to be sent based on the abnormal distribution and the historical baseline value. According to the invention, by customizing the abnormality judgment condition according to the index type and combining the historical baseline value comparison, identification and filtering of redundant alarms of'frequent historical abnormality but no essential risk at present ', the interference of invalid notifications on operation and maintenance personnel is reduced, and only alarms of'significant difference between current abnormality and historical rules' are sent, so that the operation and maintenance efficiency is improved. The operation and maintenance personnel are ensured to focus on system risks really needing to be processed, and the situation that important problems caused by alarm flooding are ignored is avoided.

1. Name of the Design Product: Thoracic Electrical Impedance Tomography Instrument. 2. Use of the Design Product: For thoracic impedance measurement, and for evaluating and studying the abnormal distribution and dynamic laws of intrapulmonary gas caused by various physiological or pathological states. 3. Design Key Points of the Design Product: The combination of shape and pattern. 4. Picture or Photograph that Best Illustrates the Design Key Points: Perspective view.

The application belongs to the field of image defect intelligent identification, and more particularly to a screen defect detection method and system based on a cooperative neural network. A trained feature relationship discrimination network can be used to obtain target defect distribution feature maps corresponding to each defect type according to target fusion defect feature maps and target defect features corresponding to each defect type, so that the distribution of each type of defect in the target feature map can be preliminarily obtained using a small number of samples of each type of defect image, and the generalization ability of detection and recognition is improved. The cosine similarity between each block of the target feature map and the defect-free feature map is calculated to obtain an abnormal distribution feature map that can accurately reflect the distribution of each abnormal block in the target feature map. Then, the overlap of the anchor boxes in the abnormal distribution feature map and the target defect distribution feature map corresponding to each defect type can be used to accurately obtain the actual detection and recognition result of the target feature map for multiple defect types in a double-channel joint determination manner.

The embodiment of the application provides a kind of merchant early warning monitoring method, system and terminal equipment based on multidimensional data fusion, belong to data fusion technical field.The method comprises: obtaining the historical operation data and user evaluation text of historical merchant, and the first abnormal data and the first abnormal distribution are obtained by carrying out abnormal detection to historical operation data;First description text is determined according to first abnormal data and first abnormal distribution;Second abnormal data and second abnormal distribution are obtained by clustering analysis to user evaluation text;Second description text is determined according to second abnormal data and second abnormal distribution;The historical correlation under historical operation data and user evaluation text is determined according to first description text and second description text;Risk prediction model is established according to historical operation data, user evaluation text in combination with historical correlation;According to risk prediction model, current operation data and current evaluation text are combined, and target merchant is early warning monitored to obtain target monitoring result.

The invention relates to an incentive identification system and method for an abnormal wafer. The method comprises the following steps: acquiring an abnormal distribution diagram of the abnormal wafer; the abnormal distribution diagram comprises a defect distribution diagram of the wafer and / or a yield distribution diagram of the wafer; comparing a feature mark of a part contacted with the wafer on the surface of the wafer when a process machine carries out the process with the abnormal distribution diagram, and finding out the feature mark with a similar area with the abnormal position in the abnormal distribution diagram and the part corresponding to the feature mark; and according to the part corresponding to the feature mark, obtaining a process machine and process steps which cause wafer abnormity in the process. According to the invention, the intelligent analysis of the abnormal wafer is realized, the automation degree is improved, and the engineering analysis efficiency is improved.

The application discloses a kind of equipment failure early warning detection methods based on digital radar holography scanning, it is related to the technical field of failure early warning, including, the abnormal propagation direction and parameter relationship transmission characteristics between the analysis secondary abnormal distribution set and primary abnormal distribution set, form abnormal evolution chain, extract abnormal structure mode of abnormal evolution chain by structure analysis, and execute fault fitting processing, generate multidimensional fault fitting set;Estimate the abnormal evolution trend of multidimensional fault fitting set, generate operation risk index set, execute abnormal evolution consistency determination to operation risk index set, generate failure early warning result set.The application deduces and constructs abnormal evolution chain by counterfactual state, realizes the explicit modeling to equipment "normal operation relationship mode", makes abnormal change from "discrete identification" to "structured evolution analysis", improves failure early warning foresight and reliability.

The invention relates to the field of stamping die control, and particularly discloses an intelligent control system and method for a precision hardware stamping die, and the method comprises the steps: drawing and preprocessing a force-stroke curve based on the stamping force and stroke data of an upper die measuring point; generating a standard force-stroke curve and a reference tolerance zone thereof based on the stable production data; calculating deviations of typical characteristics such as the peak force, the stroke position reaching the peak force, the curve slope of a plastic deformation area and the force value of a specific stroke section between each stroke-frequency curve and the standard curve; according to the characteristic deviation and the time sequence change trend, early warning is triggered, and an abnormal curve is recognized; and by analyzing the deformation consistency, the abnormal distribution range and the change mode of the abnormal curve, judging whether the abnormal source is a material problem or a mold problem. According to the method, force-stroke dynamic information is fused, early trend early warning and root diagnosis are realized, and the problems of single monitoring dimension, early warning lag and diagnosis deficiency in the prior art are solved.

The present invention relates to the field of archive management technology, and in particular to an intelligent archive management system based on RPA, the system comprising: a storage module identifies an invoice structure and performs distributed storage; a frequency setting module collects first node performance data to determine an initial detection frequency; a three-type node analysis module detects whether the three-type storage node is abnormal, adjusts the initial detection frequency according to the abnormal judgment result, or generates a second-type storage node abnormal judgment instruction; a second-type node analysis module determines whether the second-type storage node is abnormal based on the node abnormality ratio or node abnormality ratio change trend of a number of three-type storage nodes, so as to formulate a second-type node repair strategy according to the comparison result of the number of second-type abnormalities and the preset number of abnormalities, or generates a first-type storage node abnormal judgment instruction; a first-type node analysis module analyzes the abnormal distribution of a number of abnormal second-type storage nodes, so as to determine the first-type node repair strategy based on the degree of abnormality. The present invention improves the efficiency and security of archive management.

The present invention relates to the technical field of intelligent supervision. Disclosed are a distribution-network-device state simulation method and system based on fault coupling analysis. The method comprises: collecting state data of a distribution network device, and pre-processing the collected data; performing analysis and mining on the state data of the distribution network device, and extracting a device fault mode, a coupling relationship and an anomaly; providing a visualized analysis result, and providing real-time monitoring and alarming functions, such that the anomaly of the distribution network device is discovered in a timely manner; and formulating an intelligent scheduling algorithm to search for the specific location of the abnormal distribution network device. The present invention can implement decision-making and planning of anomalies, thereby facilitating the timely discovery of anomalies in distribution networks.

The invention discloses a method and a system for judging the quality of packaging test data of an integrated chip under abnormal distribution, belongs to the technical field of integrated chip detection, and solves the problem of how to judge the packaging quality of the integrated chip under the abnormal distribution. According to the test data distribution, different methods are selected to calculate the process capability index of the test data, and according to the data concentration degree and the process offset degree in the process capability index, the packaging quality grade of the integrated chip is judged; when the integrated chip packaging test data is in non-normal distribution, 1% significance level is used for filtering interference data with overlarge offset and a peak mean value is used for replacing a process mean value, so that a process capability index calculated by the integrated chip packaging test data under the non-normal distribution has higher anti-interference capability; and the reliability of judging the packaging quality of the integrated chip is higher.

The application discloses a power plant thermal system multi-parameter correlation anomaly detection and positioning method, and belongs to the technical field of industrial process state monitoring and fault diagnosis. By acquiring historical normal operation data, a multi-source fusion graph is constructed according to a physical connection relationship and statistical correlation; a graph neural network is trained by using normal working condition data to obtain a health benchmark model. Real-time data are acquired, input into the health benchmark model to calculate reconstruction error, and it is determined whether the system is abnormal. When it is determined that the system is abnormal, a real-time fine graph representing a current abnormal state is subjected to multi-stage graph coarsening, coarse graph sequences of different resolutions are generated, abnormal distribution is analyzed layer by layer to position a fault starting node from coarse to fine, and in combination with the correlation change characteristics of the fault starting node and neighbor nodes, the fault type is determined. The application solves the problem that multi-parameter correlation anomalies are difficult to quickly and accurately position.

The present application relates to the technical field of leprosy nerve injury grading evaluation, in particular to a leprosy nerve injury intelligent grading evaluation method based on high-frequency ultrasound multi-parameter fusion, which adopts high-frequency ultrasound to obtain multiple frames of nerve images, analyzes gray scale trends to construct a closed path, screens out abnormal distribution frames to generate a consistent sequence, extracts regional coverage, brightness morphology and texture arrangement features, generates change descriptions by comparing feature differences, matches preset grading rules according to description types, determines the injury level, and obtains the leprosy nerve injury intelligent grading evaluation result. The present application adaptively defines the nerve region by using the gray scale change of the image, reduces the experience dependence, ensures the spatial homology and stability of the feature comparison by screening the consistent image sequence, weakens the single index interference based on the overall trend of the change description, guarantees the coherence and distinguishability of the evaluation result under the complex background, and enhances the applicability of the grading conclusion in the dynamic observation scene.

The invention discloses a novel electric power system safety prediction and early warning method based on AI, and relates to the technical field of electronic power. Obtaining a contribution value range of the power abnormal condition to the action index associated with the power abnormal condition when the power abnormal condition exists, and taking the contribution value range as an abnormal contribution range; forming a weight coefficient of the power abnormal condition relative to the associated action index; when the real-time data of the action index exceeds the normal value range of the action index, taking the action index as an abnormal action index; obtaining an exception distribution value of the power exception condition; obtaining an abnormal prediction distribution value of the power abnormal condition; and if yes, giving out early warning. By forming the weight coefficient of the power abnormal condition for the action index, performing abnormal decomposition on the real-time data of the abnormal action index and performing joint distribution verification on the abnormal prediction distribution value of the power abnormal condition, the accuracy of abnormal prediction can be improved, and then targeted early restoration can be performed on the abnormality.

The invention relates to a method and system for checking whether zero calibration of a spiral feeder is accurate or not. The method comprises the steps that the spiral rotating speed and the corresponding discharging amount of the spiral feeder in the normal working period are obtained; performing data cleaning on the spiral rotation speed of the spiral feeder and the corresponding blanking amount data to obtain data after data cleaning; carrying out normality verification and descriptive statistics on the spiral rotation speed in the same blanking amount interval; when the statistical result of the data is in abnormal distribution, data cleaning is carried out again; when the statistical result of the data is normal distribution, determining a normal operation range of the spiral rotation speed by using a probability density function of the normal distribution; and when the spiral rotation speed of the target spiral feeder continuously exceeds the normal operation range, an abnormal calibration alarm is given out. According to the method, the normal operation range of the spiral rotation speed is defined according to the statistical rule (normal distribution interval), and the operation reliability and the production stability of the spiral feeder can be improved.

The invention discloses an apparent defect detection system and method applied to an electronic product, and relates to the technical field of defect detection.The method comprises the steps that historical detection records obtained after apparent defect detection are completed are collected, feature bit sequences are judged and extracted according to defect abnormal distribution, and defect grade evaluation is conducted; sorting defect level conditions of the feature bit sequences, and analyzing relevance among the feature bit sequences; collecting detection records after manual repair evaluation is completed, summarizing unrepairable records, and judging a defect total degree threshold which affects a repair result; collecting the detection records with the defect total degree exceeding a threshold value, analyzing the defect level of each feature bit sequence, and judging and extracting a primary bit sequence; the real-time product image corresponding to the primary bit sequence is collected, the image defect level is evaluated, the preset detection process is optimized, and the production efficiency and the quality reliability of the electronic product are improved.