609results about How to "Realize identification" patented technology

The invention relates to an unmanned aerial vehicle visual identification method and device, and the method comprises the steps: collecting a multi-modal image through a multi-modal sensing device of an unmanned aerial vehicle, and carrying out the image conversion of the multi-modal image, and obtaining an enhanced image; performing target detection on the enhanced image according to a preset target detection network to obtain a preliminary detection target; performing regional feature analysis on the preliminary detection target to obtain corresponding target regional feature information; acquiring environmental factors collected by the multi-modal sensing device, and performing environmental tracking analysis on the preliminary detection target according to the environmental factors to obtain a target tracking result; and performing multi-task analysis on the preliminary detection target through a preset recognition model in combination with the target region feature information and the target tracking result to obtain a corresponding comprehensive recognition result. The overall performance and reliability of visual identification of the unmanned aerial vehicle can be improved.

The invention aims to provide a method and a device for determining the peak area of chloride ions based on ion chromatography. The method for determining the peak area of chloride ions based on ion chromatography comprises the following steps: step 1, identifying a chromatographic peak: peak separation, peak qualitation, baseline correction, peak trailing treatment and peak overlapping treatment; and step 2, calculating the chromatographic peak area: peak definition, integration algorithm, calibration curve application, repeatability verification, impurity peak deduction and main peak area calculation. According to the method, a chromatographic peak identification and chromatographic peak area calculation mode is designed, so that the technical problem that other impurity ions may be introduced to influence chloride ion chromatographic peak identification and accurate area calculation in an existing mode of determining organic chlorine through ion chromatography is solved.

The embodiment of the invention provides a rape development period identification method and device based on multi-modal big data, and is applied to the technical field of crop identification. The method comprises the following steps: acquiring an image data set and a modal data set of rape; preprocessing the image data set and the modal data set; inputting the preprocessed image data set into a preset image feature extraction model to output image features, and inputting the preprocessed modal data set into a preset text extraction model to output semantic features; inputting the image features and the semantic features into a preset multi-modal feature fusion model to output fusion features; and inputting the fusion features into a preset developmental period identification model to output the rape developmental period. In this way, the recognition accuracy of the developmental periods such as the squaring period of the rape and the like with unobvious image features can be effectively improved, and accurate recognition of the whole developmental period of the rape is achieved.

The invention relates to the technical field of brain function imaging, in particular to a target crowd recognition method and device based on neural blood flow feature multi-level fusion, and the method comprises the steps: collecting brain nerve blood flow signals of normal and target crowds by using synchronous EEG-fNIRS equipment, preprocessing and extracting single-mode features, and obtaining a single-mode feature set; a multi-frequency local nerve and blood vessel coupling analysis method is used for calculating nerve and blood vessel coupling strength characteristics, a space co-positioning method is used for unifying nerve and blood flow modal space information, the position precision is improved, and data layer fusion is achieved; and integrating and analyzing different modal feature space relationships by using a Yeo7 brain template to realize feature layer fusion, and finally training various machine learning classifiers by using significant difference features to realize identification of a target population. Therefore, the problems that in the prior art, macroscopic signal correlation is paid more attention to, so that analysis of the synergistic effect of cross-modal space information in a specific brain region is deficient, and deep understanding of a neurovascular function mechanism and brain region role revelation are restricted are solved.

The invention belongs to the technical field of radar interference identification, and discloses a radar interference identification method based on LSTM and a variational auto-encoder. According to the method, a joint network of the LSTM and the variational auto-encoder is constructed to encode the time sequence signal to obtain a corresponding feature vector, then the feature vectors of the emission signal and the echo signal are spliced to obtain a combined vector capable of representing the interference type, and then classification learning is performed on the combined vector through the convolutional neural network, so that the interference type can be represented. And identification of interference signals is realized. By means of the capability of the LSTM network to solve the long-term dependence problem and the advantages of the variational auto-encoder in dimension reduction and feature extraction, time sequence data can be effectively processed, and complex distribution of data can be learned; and meanwhile, vectors before and after the interference process are spliced to represent a change relation before and after the signal, so that interference pattern identification under any change can be adapted.

The invention discloses an autonomous cognitive algorithm of a patrol robot hidden area based on a 3D laser radar, and relates to the technical field of artificial intelligence. The invention provides an autonomous cognitive algorithm for a patrol robot hidden area based on a 3D laser radar. The autonomous cognitive algorithm comprises the following steps: acquiring a non-ground point cloud by adopting a ground point cloud filtering algorithm; for the obtained non-ground point clouds, dividing concentric circle regions according to Euclidean distances, and performing Euclidean clustering on the point clouds in the concentric circle regions to obtain clustered point cloud clusters; according to the point cloud cluster, calculating an axis alignment bounding box, and obtaining a hidden point coordinate; and according to the point cloud cluster, calculating a directional bounding box, and calculating coordinates of a shielding point in a sight line shielding mode, thereby automatically calculating and obtaining a shielding angle. Therefore, by the adoption of the method, the hidden area is reasonably selected according to the angle information in the patrol process through the 3D laser radar data, the hiding effect is achieved, the patrol task is better completed, and high robustness and practicability are achieved.

The invention discloses an underwater target identification method for an underwater intelligent robot, which comprises the following steps: firstly, enhancing the definition and color restoration effect of an underwater image through image enhancement processing, and then utilizing an underwater target identification network model which is designed in a targeted manner and is improved based on a feature fusion attention network framework to identify the underwater target. A new defogging target detection model is obtained, a clear fog-free image is output in an end-to-end mode, and recognition of an underwater target is achieved; according to the method, a good foundation is laid for the underwater intelligent robot to carry out underwater tasks, and an underwater target recognition network model improved based on a feature fusion attention network framework is deployed in a detection module of the underwater intelligent robot, so that the feasibility of deploying an underwater target recognition algorithm in mobile terminals such as the underwater intelligent robot is improved; the accuracy and real-time performance of underwater target recognition in an underwater intelligent robot are improved, and the universality and practicability of an underwater target recognition algorithm based on a neural network in mobile terminal equipment are improved.

The invention provides a time-varying structure vibration signal decomposition method based on experience multiple synchronous extraction, and the method comprises the steps: carrying out the frequency spectrum segmentation through searching a sudden change point of the internal and external energy of a characteristic frequency band in a signal time-frequency spectrum, so as to complete the time-varying signal decomposition: firstly, obtaining the time-frequency spectrum of a signal through the multiple synchronous extraction transformation; summing the time-frequency coefficients of all moments at each frequency position to obtain time-frequency total energy; searching an energy boundary by using an energy segmentation operator, and performing Fourier transform on the signal to obtain a signal frequency spectrum; then, combining a zero-phase filter and an energy boundary position, and extracting a frequency band corresponding to the single-mode signal; and finally, inverse Fourier transform is carried out on each section of frequency spectrum, a single-mode signal is reconstructed, and modal decomposition of the time-varying structure is realized.

The invention discloses a frame weight-based dynamic micro-expression recognition method and a related device, and relates to the technical field of image processing, and the method comprises the steps: obtaining a to-be-recognized micro-expression image sequence, inputting the to-be-recognized micro-expression image sequence into a trained dynamic micro-expression recognition model, and obtaining a micro-expression recognition result; the dynamic micro-expression recognition model comprises an initial convolution module, a frame attention module, a dynamic and static combination module and a time sequence processing module. According to the method, the frame attention module is adopted in the dynamic micro-expression recognition model, the frame attention weight can be calculated while the channel attention weight is calculated, recognition of the key frame in the dynamic micro-expression image sequence is achieved, and by adopting the dynamic and static combination module, the recognition accuracy of the key frame in the dynamic micro-expression image sequence is improved. Dynamic information between the feature maps is added on the basis of static feature map processing, so that the model is more sensitive to change details of dynamic micro-expressions, and the micro-expression recognition accuracy is greatly improved.

The invention relates to a method and device for confirming the number of battery packs in a battery system, the battery system, a computer readable storage medium and a computer program product. The method comprises the following steps: identifying a bus voltage of an accessed target battery system; the reference voltage of a single battery pack in the target battery system is obtained, and the reference voltage is the product of a preset static voltage corresponding to a single battery cell in the battery pack and the number of battery cells contained in the battery pack in the state of the lowest residual electric quantity; determining a voltage ratio of the bus voltage to the reference voltage; and rounding the voltage ratio to obtain the number of battery packs in the target battery system. According to the method, the total bus voltage of the static battery system is directly identified, and then the number of the stacked battery packs in the system is determined through the ratio of the bus voltage to the reference voltage, so that the identification of the number of the battery packs can be realized without charging processing, and the efficiency of the battery pack number identification processing process can be effectively improved.

The invention relates to a railway tunnel portal safety state detection method, system, device, medium and program, and the method comprises the steps: carrying out the two-dimensional data collection of a railway tunnel portal to obtain a tunnel portal image, and carrying out the three-dimensional data collection of the railway tunnel portal to obtain a tunnel portal point cloud; performing multi-layer target feature detection and two-dimensional target labeling on the tunnel portal image to obtain a two-dimensional detection result; performing point cloud denoising and spatial voting clustering on the tunnel portal point cloud to obtain a primary point cloud class group; performing bounding box detection and three-dimensional target labeling on the tunnel portal point cloud according to the primary point cloud class group to obtain a three-dimensional detection result; and performing detection target alignment and scene modeling on the two-dimensional detection result and the three-dimensional detection result to obtain a standard detection result, and performing safety analysis on the standard detection result to obtain the tunnel portal safety state, thereby improving the efficiency of the tunnel portal safety state.

The invention provides a user call emotion real-time identification method and system based on ASR. The ASR-based user call emotion real-time recognition method comprises the steps of collecting voice data in a user call process in real time, and performing segmentation processing on the voice data by using a comprehensive voice segmentation strategy in combination with a first segmentation coefficient and a second segmentation coefficient to obtain a plurality of to-be-recognized voice data segments; performing text conversion on the to-be-recognized voice data segment by using an ASR voice automatic recognition system to obtain text data corresponding to the to-be-recognized voice data segment; preprocessing the text data in combination with a to-be-recognized voice data segment to obtain preprocessed text data; and inputting the preprocessed text data into a trained deep learning model for emotion analysis, and obtaining an emotion recognition result corresponding to the preprocessed text data. The system comprises modules corresponding to the steps of the method.

The invention discloses an intelligent fund analysis system for a specific gang behavior of a block chain, and relates to the technical field of block chains, and the system comprises a data collection unit which is used for obtaining data points based on data on all nodes of the block chain; the data fusion unit is used for obtaining first data based on the data points; the label unit is used for obtaining second data based on the label system and the first data; the address anomaly analysis unit is used for obtaining transaction data based on the second data and judging whether the to-be-identified address has a specific gang behavior based on the transaction data; the graph analysis unit is used for constructing a fund flow graph, extracting graph features of the fund flow graph and obtaining key data based on the graph features; the comprehensive analysis unit is used for pre-warning potential risk behaviors based on graph features; and the result unit is used for obtaining an analysis result. The problems that in the prior art, specific gang behavior fund flow based on the block chain cannot be accurately identified and predicted, and block chain characteristics and constantly changing supervision requirements cannot be adapted can be solved.

The invention discloses a liquid storage type micro-fluidic chip and a working method thereof. The chip comprises a chip body, a sealing film and gum, the sealing film and the back adhesive are respectively bonded on the front surface and the back surface of the chip body; a plurality of solution storage cavities are formed in the chip body, outflow runners at the bottoms of the solution storage cavities are connected with corresponding first runners, sealing partition pieces are arranged on the first runners, the outflow runners and the corresponding first runners are blocked through the sealing partition pieces, and the first runners are communicated with second runners through corresponding gum runners. The whole experiment process does not need to be externally connected with a reagent bottle, the purposes of instant detection, preparation and sample separation of the micro-fluidic chip can be achieved, the detection time is greatly shortened, and the experiment steps are simplified. Waste liquid generated in the experiment process flows into the waste liquid cavity, and cross contamination of a laboratory is avoided. The chip provided by the invention can be used for identifying the species of pathogenic microorganisms.

The invention relates to the technical field of arch dam deformation monitoring data processing, in particular to an arch dam deformation monitoring data preprocessing method and system, and the method comprises the steps: collecting arch dam deformation monitoring data; constructing a track matrix for the original deformation monitoring data sequence, and performing singular value decomposition on the track matrix to obtain a feature value and a feature vector; selecting a principal component according to the contribution rate of the characteristic value, and reconstructing a signal; the reconstructed signals are analyzed, and gross errors are identified and eliminated; dividing the deformation monitoring data without the gross error into a training set and a test set; constructing a neural network model, and training the neural network model by adopting the training set; and filling missing data in the test set by using the trained neural network model. According to the invention, monitoring data gross error identification and monitoring data missing processing are realized.

The invention discloses an intelligent laser weeding dynamic fine positioning method and system based on machine vision, and the method comprises the steps: carrying out the target detection based on an image collected by a coarse positioning camera, and carrying out the coarse positioning of a detection frame of weeds; carrying out optical flow field calculation based on an image acquired by the coarse positioning camera to obtain a displacement offset; correcting the position of a coarse positioning detection frame in real time based on the displacement offset; controlling a fine positioning camera to carry out image acquisition according to the corrected position; and calculating the center position of the weed rhizome based on the acquired image. The accurate weed center pixel position is obtained through combination of coarse positioning and fine positioning, the motion direction and distance of a machine are predicted through an optical flow method, dynamic and accurate positioning is achieved, and the problem that in existing laser weeding, the weed visual recognition precision is not high is solved.

The invention discloses a bearing vibration signal processing model construction method, a model and a signal classification method, which are used for constructing time domain and frequency domain loss functions by calculating a plurality of feature vector distances so as to train an encoder. According to the method, by minimizing the difference value of different distances in the time domain and frequency domain loss functions, an encoder can be guided to extract similar feature representations in signals, so that the similar feature representations are gathered in a feature space, and dissimilar features are separated. When a trained encoder (a convolutional neural network with an attention mechanism) processes a vibration signal of an unknown fault type, some feature representations in the signal are different from those of other signals, so that the feature representations are gathered in a feature space and are far away from the feature representations of other signals, and the fault detection accuracy is improved. Therefore, the recognition of the vibration signal of the unknown fault type is effectively realized.

The invention provides a track inspection robot abnormal sound detection method based on a convolution auto-encoder, and the method comprises the steps: firstly collecting sound data of the normal work of coal mine transportation equipment in an inspection region through sound pickup equipment carried on a track inspection robot, and carrying out the environment noise reduction of the sound data through a noise reduction algorithm; secondly, extracting Mel-frequency cepstral coefficient features of the sound of the coal mine transportation equipment through a feature extraction algorithm, inputting normal audio features into a convolutional auto-encoder network for training to obtain reconstructed audio features, and calculating a mean square loss function of the input normal audio features and the reconstructed normal audio features to obtain a reconstruction threshold value; and finally, calculating a mean square loss function of other audio features input into the auto-encoder network, and detecting the abnormal sound of the coal mine transportation equipment in a mode of judging whether the mean square loss function exceeds a reconstruction threshold value or not. According to the detection method, the problem that enough abnormal samples are difficult to accumulate in a short time due to the fact that the abnormal sound of the coal mine transportation equipment has occasionality and diversity can be solved, and the abnormal sound of the coal mine transportation equipment can be detected by the track inspection robot under the condition that the abnormal samples are lacked; and the production safety problem caused by the abnormity of the coal mine transportation equipment in the coal mine production can be early warned in time.

The invention discloses a ground penetrating radar tunnel lining disease identification method and device, and the method comprises the steps: carrying out the amplitude-frequency feature analysis of tunnel lining data collected by a ground penetrating radar, and obtaining the corresponding amplitude-frequency feature data; generating a radar reflection signal profile map according to the tunnel lining data, and performing disease marking on the radar reflection signal profile map to obtain a profile map after disease marking; inputting the amplitude-frequency characteristic data into a characteristic deconstruction analysis model, and outputting a first disease identification result; inputting the section map after the disease marking into a section map disease identification model, and outputting a second disease identification result; and carrying out weighted average on the first disease identification result and the second disease identification result to generate a final tunnel lining disease identification result. According to the invention, the identification precision and identification efficiency of tunnel lining diseases can be improved.

The embodiment of the invention discloses a steel bar binding point panoramic recognition method and system and electronic equipment. The method comprises the following steps: acquiring a video stream shot by an industrial camera mounted on a portal frame under a moving condition to extract images and obtain a plurality of camera view images; fusing and splicing the plurality of camera view images to obtain a scene panoramic image; performing cross intersection detection on the scene panoramic image to obtain intersection information; wherein the cross point information comprises a target frame of a bound cross point and a target frame of an unbound cross point; and the cross point information is sent to a binding mechanical arm, so that the binding mechanical arm determines the coordinates of the cross cross points which are not bound, and binding is carried out. By implementing the method provided by the embodiment of the invention, the panoramic view angle image of steel bar binding can be identified, so that all binding points can be effectively identified and detected, and potential safety risks are avoided.

The invention belongs to the technical field of electromagnetic interference elimination, and discloses a method and system for processing electromagnetic leakage interference of an electromagnetic vibration table, and the method comprises the steps: starting a signal collection and analysis module, monitoring an electromagnetic signal generated in the operation process of the electromagnetic vibration table in real time, and recognizing an interference frequency range; a proper filter is selected or designed based on the identified frequency range through a filter selection and design module; a filter processing unit is utilized to install filters on necessary power lines and signal lines, and interference suppression measures are executed; the test and optimization module is operated, the filtering effect is evaluated, the configuration of the filter is adjusted according to needs, and the electromagnetic interference suppression effect is optimized; and the monitoring and adjusting module is continuously executed to adapt to the change of an experimental environment, so that long-term effective electromagnetic interference control is ensured. The invention provides a solution which is more accurate, effective and flexible compared with the traditional physical shielding, and is suitable for the occasion of carrying out high-precision vibration test on the magnetic sensitive sample.

The embodiment of the invention discloses an enterprise IP asset identification method and electronic equipment, and the method comprises the steps: constructing an atlas through website data in combination with historical data stored in PDNS data, determining a target search path with a target domain name corresponding to a to-be-identified enterprise as a source node and an IP as a termination node based on the constructed atlas, and carrying out the identification of the enterprise IP assets through the target search path. And inputting the target feature vector of each target node in the target search path into an asset sequence classification model based on a pre-trained asset sequence classification model according to the first feature vector of each node in a pre-stored atlas, and determining whether the IP and the target domain name belong to the same enterprise asset, thereby realizing identification of the enterprise IP asset. Therefore, various data historical source IPs in a complex network environment are effectively integrated, and the comprehensiveness and accuracy of identifying IP assets are ensured.

The invention relates to an overhead optical fiber sensor event identification method and system, and belongs to the field of cable online monitoring, and the method comprises the steps: obtaining a vibration signal at a corresponding position through an optical fiber sensor; decomposing the vibration signal through a wavelet packet decomposition method to obtain a plurality of signal components of different frequency bands, weighting each signal component, and performing inverse transformation to obtain a reconstructed signal; extracting a vibration energy value at each sampling point from the reconstructed signal to generate a thermodynamic diagram; and carrying out image identification and classification on the thermodynamic diagram to obtain a sensor event and an event category. Vibration signals are processed through a wavelet packet decomposition method to obtain signal components of different frequency bands, environment noise in a low-frequency band is weakened through weighting to highlight event noise in a high-frequency band, a reconstructed signal is obtained through inverse transformation, a thermodynamic diagram is generated by using the reconstructed signal, and the performance of the thermodynamic diagram is improved. The area corresponding to the event in the thermodynamic diagram is identified and analyzed through the image identification algorithm, so that event identification is realized, a threshold value does not need to be set, the adaptability is better, and the efficiency is higher.

The invention discloses a pattern recognition method and device for medical gene sequencing data, and belongs to the technical field of data processing.The method comprises the steps that a target gene sequence obtained through sequencing is preprocessed, a self-adaptive scoring matrix is constructed, and site scores are dynamically adjusted through neural network training; by means of the self-adaptive scoring matrix, a sliding window is adopted to scan sequence differences, and a Gaussian mixture model is combined for analysis, and a variation region information file is generated; reading a variation region information file, constructing a base conversion matrix for sequence comparison, and generating a variation classification result by analyzing a comparison path; and based on a variation classification result, performing standardization processing on the variation region feature vector, and generating a mutation mode type through an improved clustering algorithm. Through the scheme of the invention, the filtering quality of the sequence data can be improved, and the capability of processing a complex variation mode is enhanced.

The invention relates to the technical field of image processing, and discloses a container loading and unloading positioning identification method based on machine vision. The method comprises the following steps: firstly, obtaining a to-be-processed container image, and carrying out noise reduction processing, binarization processing and edge detection on the to-be-processed container image to obtain a processed container image; secondly, HOG features of the processed container image are extracted, and an HOG feature vector set is obtained; a final SVM classifier is obtained through training, the HOG feature vector set is input into the final SVM classifier for classification, and container recognition is completed. And after container identification is completed, obtaining a container top image, processing the container top image to obtain a container corner fitting lockhole image, and positioning the position of the container corner fitting lockhole by using a convolutional neural network to realize container loading and unloading identification and positioning. According to the method, the container image is processed, the purpose of container loading and unloading positioning identification is achieved, and the method is accurate and objective.

The invention relates to the technical field of communication adjustment, and discloses a communication system and method for diabetes outpatient service user management, and the system comprises an acquisition module which is configured to acquire real-time communication data of a plurality of communication devices in a diabetes management system; and the judgment module is configured to obtain network parameters of each communication device, and is further configured to determine a communication fluctuation value according to a relationship among the historical communication data. The prediction module is configured to predict a communication performance change trend of each communication device in a preset time period based on the long short-term memory network, and generate a prediction threshold. The control module is configured to dynamically adjust the communication path according to the relation between the real-time communication fluctuation value and the prediction threshold value. The encryption module is configured to dynamically encrypt the adjusted communication path and data. The communication efficiency of diabetes outpatient service user management can be improved by collecting the communication data in real time, predicting the performance change trend, dynamically adjusting the communication path and encrypting the protection data.

The invention provides an online track point recognition method based on a self-supervised learning model, and belongs to the technical field of deep learning, and the method comprises the steps: obtaining unlabeled sample data in a training set, carrying out the pre-training of the self-supervised learning model, and obtaining a mask encoder; transferring parameters of a mask encoder in the self-supervised learning model to an encoder of an end-to-end classification model, establishing an initial end-to-end classification model by connecting a classification head, training the initial end-to-end classification model through a small amount of labeled sample data in a training set, and obtaining the end-to-end classification model; and an encoder in the end-to-end classification model performs feature extraction on the parameter information of the online track points acquired in a preset time period, and a track point identification result is acquired through classification head prediction. According to the method, on the basis of the self-supervised learning model and the end-to-end classification model, online track points can be recognized, and complex track points are recognized while track point recognition efficiency is improved.

The invention provides an igneous rock fluid identification method based on a CXDF algorithm, and the method specifically comprises the following modules: S1, data preprocessing: obtaining core and logging data of a research area, and sorting the data into a table file in an xlsx format; performing logarithm processing on the specified logging data; performing normalization processing on the logging data; 78% of the preprocessed data is a training data set, and 22% of the preprocessed data is a verification data set; dividing a training data set label into two parts, namely multi-hydrocarbon and low-hydrocarbon; s2, model establishment: constructing a fluid identification model based on a CXDF algorithm and other artificial intelligence algorithms; constructing a simulation well based on the verification data set; identifying the simulated well according to the fluid identification model; s3, model optimization: performing model evaluation and model comparison based on a table, a radar map and a histogram; and determining an optimal fluid identification model. According to the method, rebalance processing is carried out on the data set based on an adaptive multi-target group cross optimization algorithm, and recognition of the complex lithologic igneous rock reservoir fluid is realized in combination with an integrated learning model.

The invention provides a lithology identification method and device based on a two-channel convolutional neural network, and relates to the technical field of information processing, and the method comprises the steps: obtaining a hyperspectral image sample and a multi-temporal multispectral image sample; preprocessing the hyperspectral image sample and the multi-temporal multi-spectral image sample, respectively inputting the preprocessed hyperspectral image sample and the preprocessed multi-temporal multi-spectral image sample into a one-dimensional convolution model and a two-dimensional convolution model, and outputting to obtain a target spectral feature and a target spatial feature; fusing the target spectral feature and the target spatial feature to obtain a fused feature, and training a preset fusion model based on the fused feature to obtain a target fusion model; and outputting a to-be-identified image to the target fusion model, and determining the lithology type corresponding to the to-be-identified image. According to the method, different features are learned by using the one-dimensional convolution model and the two-dimensional convolution model respectively, and the output features of two different channels are fused, so that the lithology of the remote sensing image is identified, and the accuracy and reliability of an identification result are improved.

The invention relates to the technical field of visual inspection, in particular to a stamping part defect detection method based on vision, which comprises the following steps: S1, shooting setting; s2, synchronizing video streams; s3, acquiring a part bounding box and a hole position bounding box; s4, associating part bounding boxes belonging to the same stamping part among the frame part images, and associating hole position bounding boxes belonging to the same screw hole position; s5, correspondingly associating the bounding boxes of the target parts belonging to the same stamping part in the part image; s6, when a hole position boundary frame matched with the view angle template image of the corresponding view angle exists in any target detection image under any view angle, determining that a screw hole position exists; and when the matched hole position boundary frame is not found in the target detection image under the corresponding visual angle in the visual angle template images of all the visual angles, the screw hole position is missing. The stamping part defect detection device can be used for defect detection of stamping parts in the conveying process, the consistency of detection results is improved, and the detection efficiency is improved.