24results about How to "Suppress misjudgment" patented technology

The invention relates to the technical field of industrial vision multispectral imaging detection, in particular to a defect detection method and system based on multispectral fusion imaging, and the method comprises the steps: synchronously collecting a multi-channel spectral image of the surface of an object to be detected, and generating an enhanced feature map fusing multispectral information; identifying and preliminarily marking a suspected defect region by applying a region segmentation algorithm based on anomaly detection; extracting a multi-dimensional spectral response curve of each region in each original spectral channel, and constructing a spectral feature vector; matching the vector with a standard template of a pre-established defect spectral feature database, and classifying and confirming defect types; and carrying out contour refined analysis, calculating the geometric dimension and position of the defect, and generating a structured detection report. According to the method, the sensitivity of defect detection and the classification accuracy are improved through accurate matching of multispectral fusion and spectral features, and automatic and high-precision defect identification is realized.

The application discloses a beacon user selection method for refining VR boundary in an XL-MIMO system, and comprises the following steps: selecting part of users from a candidate user set A to form an initial beacon user set B , obtaining the estimated value of the VR label of the rest of the users in B according to the distance relationship; dividing all the users in A into several non-overlapping subsets according to the VR label, so that each subset corresponds to an estimated VR region, calculating the coordinate center point of each VR region and the distance between the beacon point in the region and the center point; if the distance between a user in B and any VR region center point is less than A , and the estimated value of the VR label of the user is different from the VR label of the region, then the user can be selected as a beacon user for refining the VR boundary and added to B ; the above process is repeated until all the beacon users are selected. The application can effectively solve the problem of low VR recognition efficiency caused by too few boundary beacon users in the XL-MIMO system, and make the distribution structure of the beacon users more reasonable.

The invention provides a long text-oriented analysis checking method, device and equipment and a storage medium. The checking efficiency and accuracy are remarkably improved through intelligent fragmentation and mixed scheduling. The core of the method is that according to semantic relevance among text fragments, fragments with reference or dependency are distributed to serial auditing nodes for deep and coherent context verification, independent fragments are distributed to parallel auditing nodes for concurrent processing, and therefore the problem of long text semantic segmentation is solved, and meanwhile the content of the text fragments is improved. And the auditing time is greatly shortened from the hour level to the minute level. Besides, the output behavior of the large language model is restrained by loading a predefined structured instruction set in the auditing node, so that'illusion 'and misjudgment generated when the model processes complex tables and nested terms are effectively inhibited, and the reliability of a checking result is remarkably improved. The method has flexibility and expandability, and is suitable for automatic and intelligent checking scenes of various professional long texts such as contracts and bid invitation files.

The application discloses an immersive neural rehabilitation training method based on virtual reality, and particularly relates to the technical field of neural rehabilitation training, and comprises the following steps: acquiring action data, tactile data, proprioceptive data and gaze data when a trainee performs a target action, performing uniform time sequence alignment on the data in a virtual reality environment, reorganizing according to a starting section, a pushing section and a reaching section of the target action, and outputting a target action response sequence; performing segmented comparison on the differences of multi-source responses of the target action under the conditions of visual weakening, tactile weakening and proprioceptive weakening to determine a perception dependency relationship, rearranging the prompt items and feedback items in the virtual reality scene according to the perception dependency relationship, and then performing recovery checking on the rearranged results.

This invention discloses a method and system for anomaly identification in multi-source monitoring data of steel structures, belonging to the field of steel structure monitoring data processing technology. The method includes acquiring multi-source monitoring data of steel structures and unifying time-scale alignment, generating a preprocessed dataset containing working condition features and quality markers. Based on the working condition features, the physical baseline prediction value is calculated; based on the preprocessed dataset, the data baseline prediction value is calculated, and dual residuals are obtained. The physical residual strength is calculated based on the dual residuals; and the residual inconsistency degree and sliding window adaptive threshold are calculated to output anomalies. This invention constructs a stable input through unified time-scale and quality marker constraints, generates physical and data dual baselines in parallel, and outputs dual residuals. Online discrimination is completed using correlation robustness strength + residual divergence adaptive weight + sliding window threshold, achieving an interpretable, traceable, and more robust anomaly identification process for long-term monitoring under working condition drift.

The application provides a workbench production line control method and system based on environment state analysis, comprising the following steps: S1: arranging multiple environment sensors at each workbench, collecting multiple environment elements through the environment sensors, and uniformly modeling the collected environment data to obtain an environment data model; S2: performing feature extraction on the environment data model to extract state feature parameters corresponding to each environment element; S3: performing fusion calculation on the multiple environment elements according to the state feature parameters to obtain a fusion risk parameter representing the risk level of the current work station; S4: calculating the evolution trend of the risk level with time according to the environment data model and the fusion risk parameter to obtain a risk evolution result in a prediction period; S5: determining a corresponding risk grade according to the risk evolution result; and S6: triggering a corresponding control strategy according to the risk grade to control the running state of the workbench equipment, so that the safety performance of the workbench production line can be provided.

The application discloses a centrifugal pump off-design condition diagnosis method and system based on a vibration signal, and the method comprises the following steps: collecting vibration signals of a centrifugal pump under multiple flow conditions, and constructing a vibration sample sequence with a working condition label; reconstructing the vibration signal by using a variational mode decomposition (VMD); performing a continuous wavelet transform (CWT) on the reconstructed signal, and mapping a one-dimensional vibration signal into a two-dimensional time-frequency graph; constructing a Light-CBAM-CNN diagnosis model, and training the same; inputting the vibration signal collected in real time into the trained diagnosis model after pretreatment, VMD and CWT, outputting a working condition category and a diagnosis confidence, and outputting an "uncertain" label and triggering a retest mechanism or an alarm when the confidence is lower than a threshold value. The application can realize online diagnosis of high-energy consumption off-design conditions of the centrifugal pump only by relying on the shell vibration signal, and has the advantages of small modification workload, low deployment cost, high identification reliability and strong energy-saving scheduling guiding significance.

The invention discloses a method for judging weak vital signs based on an infrared geminate transistor illumination technology, and relates to the technical field of vital sign detection, the method is applied to a detection device which can be worn on a remains finger without heart rate and body temperature and without touching vital signs, and the weak vital signs are judged on the basis of an optical signal after interaction of infrared light and finger biological tissue. The method realizes discrimination of weak vital signs and automatically triggers alarm when weak life states such as false death are identified, and comprises the following steps: collecting and decomposing a mixed optical signal to obtain direct current and alternating current components; screening a preliminary effective signal segment based on the time domain stability of the alternating-current component signal, and calculating a microcirculation physiological coupling coefficient of the preliminary effective signal segment to discriminate a target signal segment; calculating a normalized cross-correlation function peak value of the alternating-current component signals in the target signal segment as cross-correlation peak value intensity, and calculating a steady-state volume index according to the average value of the direct-current component signals; and when the cross-correlation peak intensity and the steady-state volume index both meet a preset threshold condition, determining that weak vital signs exist.

This invention discloses an automated testing method and system for the front panel of a multimedia terminal based on deep learning, relating to the fields of machine vision and industrial inspection technology. It includes: issuing preset display commands to the multimedia terminal under test via an industrial control terminal; acquiring front panel images after the terminal completes its display response; and locating small targets such as digital tubes and LEDs based on an improved target detection model. Combining region cropping and content recognition models, it identifies the displayed digits on the digital tubes and the colors and on / off states of the LEDs, generating structured recognition results based on preset state judgment rules. A multi-frame voting fusion mechanism improves detection stability, and a re-shooting and anomaly handling process is triggered based on recognition confidence. Abnormal samples are incorporated into the incremental training system, and model versions are tracked and managed to achieve continuous optimization of the detection model. The detection results are linked with the production execution system to achieve centralized storage and quality traceability of the detection data.

The application relates to the technical field of water flow structure detection, and discloses a detection method and system for the water flow structure of an indoor water tank sand wave surface, which comprises the following steps: arranging a sand wave surface and a test sand bed, arranging a camera system and a flow velocity collection probe, collecting image data for timestamp alignment, performing image recognition to extract sand wave contours and calculate sand wave state parameters, performing wavelet denoising and reconstruction on the flow velocity sequence of measuring points, performing frequency domain analysis to obtain energy spectrum distribution and calculate energy indexes, determining a backflow candidate area and a boundary position, performing energy criterion verification on the energy spectrum distribution and the energy indexes corresponding to the flow velocity sequence of the measuring points in the backflow candidate area, merging the measuring point grids that pass the verification according to connectivity to generate a water flow structure partition map and determine water flow structure characteristics, and establishing a response relationship model based on the water flow structure characteristics and the sand wave state parameters. The application improves the recognition robustness of turbulent and vortex structures and improves the accuracy of sand wave surface water flow structure detection.

The application discloses a kind of isomeric computing power net service zero-interruption endogenous security dynamic immunity method and system, belong to network security technical field.The method includes using exposure surface dynamic hopping and isomeric voting active avoidance sensing cooperative mechanism, periodically or event triggered controlled change is carried out to service interface, access entrance and resource address, combined with the isomeric parallel processing of multiple hardware and multiple models, multiple discriminant is output for the same request, and trusted result is obtained by weighted voting;According to abnormal score and uncertainty evaluation, the hopping intensity and access threshold are adaptively adjusted;Under the guarantee of session and state double writing, seamless migration is completed, and zero-interruption switching is realized.The method can significantly improve the resistance of computing power network to scanning, probing and exploitation attacks without changing the upper layer business logic, while controlling resource overhead.

The present application relates to the technical field of tire rubber material performance detection, and particularly relates to a method and system for predicting tire wet ground grip performance based on tread rubber material indexes. Six indexes are selected: A1 (low temperature tan delta area), A2 (E" peak value and peak temperature T*), B1 (small strain G" area), C1 (surface energy polarity component), C2 (retreating contact angle), and D1 (0 DEG C storage modulus), which are subjected to minimum-maximum standardization and fixed weight synthesis WGI, and the weight correction of A2 in the 0-8 DEG C peak temperature window is implemented; then linear calibration is established with the whole tire wet grip index G obtained according to ISO 23671, and the predicted G and grade (A / B / C) are output. The method is universal in equipment and simple in process, can quickly distinguish the pros and cons of the formula on a small sample scale in the laboratory, reduce road test, and improve batch consistency.

The invention provides a mobile phone holding state recognition method based on posture and pressing information, and belongs to the technical field of mobile phone holding state recognizing.The method comprises the steps that touch information of a user is collected by installing a capacitive touch button; a carrier coordinate system and a navigation coordinate system are defined, user posture information is updated and corrected in combination with the angular velocity measured by the gyroscope, and posture and pressing information is obtained by extracting Euler angles; the adjacent capacitive touch buttons are fused, and the posture and pressing information is encoded in a segmented mode, so that the posture and pressing information subjected to dimension reduction is obtained; and constructing a classifier for voting based on a preset specific holding state category to obtain a predicted holding state, and smoothing the predicted holding state by using a sliding window and sequential logic to obtain a real mobile phone holding state. According to the invention, the problems of screen mistaken rotation, privacy disclosure and large resource consumption in the lying scene of the existing mobile phone holding state identification method are solved.

This invention relates to the field of surgical assessment technology, specifically to an AI-assisted laparoscopic ileosigmoid anastomosis constipation treatment assessment system. The invention extracts the anastomosis surgical area; obtains lesion characteristic based on the positional relationship and grayscale change of each pixel within the anastomosis surgical area relative to the area center, and extracts the lesion area; based on the grayscale change of the lesion area and non-lesion area in the anastomosis surgical area in preoperative and postoperative intestinal images, as well as the smoothness of the lesion area's edge, the completeness of lesion resection is obtained, and combined with the stretching degree of the non-lesion area during surgery, the anastomosis tightness is obtained; the anastomosis tightness is assessed based on the anastomosis tightness. This invention quantifies the histological and geometric changes directly caused by the surgical actions of resection and anastomosis, more realistically reflecting the biomechanical seal of the anastomosis and improving the reliability of the assessment results.

This invention relates to the field of computer vision technology and discloses an AI-synthesized video detection method and system based on multi-agent collaboration. The method includes: acquiring the video to be detected and sampling keyframes; understanding the keyframes based on a visual language model and generating a forgery hypothesis for the current round based on the generated global semantic context; generating a spatial-temporal routing query; performing target localization and segmentation on the video to be detected based on the spatial-temporal routing query; extracting structured physical evidence from the target region video; performing evidence-constrained reasoning to obtain a preliminary judgment result and confidence level; performing consistency checks on the execution trajectory to obtain a reliability score and feedback information, and outputting the final authenticity judgment result. This invention improves the reliability, generalization ability, and traceability of AI-synthesized video detection by combining contextual modeling, hypothesis-driven routing, multi-dimensional physical evidence collection, evidence-constrained reasoning, and closed-loop reflective verification.

The invention provides a substrate processing apparatus, a determination method, a method of manufacturing a semiconductor device, and a program product, wherein the possibility of misdetermination is suppressed by determining the presence or absence of reproducibility of substrate processing according to a plurality of determination items. The substrate processing apparatus includes: a substrate processing unit that processes a substrate; a gas supply unit that supplies a gas for processing the substrate to the substrate processing unit; an acquisition unit that performs a predefined number of times of processing of the substrate, and acquires a measurement value of a predefined specific section in each of the processing of the substrate; a determination condition setting unit that sets a threshold value for each of a plurality of predefined determination items on the basis of the measurement values acquired by the acquisition unit; and a determination unit that determines whether or not the acquired measurement value is included in the range of the threshold value in each of the determination items during the processing of the substrate for which the threshold value has been set, and determines whether or not there is reproducibility in the processing of the substrate.

The invention discloses a thermal imager target detection method based on deep learning. The method comprises the following steps: S1, acquiring a thermal imaging image sequence output by a thermal imager; s2, initializing radiation correction parameters for each frame of thermal imaging image and executing radiation correction to generate a corrected thermal imaging image; s3, inputting an improved TTFNet model, calculating a center thermodynamic diagram and a target scale regression diagram, and generating a candidate target set; s4, extracting a detection structure quantity from the candidate target and executing motion compensation alignment between adjacent frames; s5, calculating detection consistency measurement and updating radiation correction parameters; s6, constructing a working state vector, and executing threshold mapping screening on the candidate target; and S7, calculating a false detection potential energy field based on the historical frame candidate target screening set, executing space suppression on the center thermodynamic diagram, and outputting a target detection result. According to the invention, the accuracy and reliability of target detection of the thermal imager are improved.

The present application relates to the technical field of remote sensing image processing and deep learning change detection, and discloses a remote sensing cultivated land change detection method. The method maps the front and rear time phase remote sensing images into feature vectors embedded with position coding, extracts features by using a 24-layer visual Transformer with low-rank matrix fine-tuning, selects the 12th, 16th, 20th and 24th layer features, obtains double time phase fusion features through a space-time interaction unit composed of a global module and a Mamba2 time sequence module in parallel, channel attention and cross-level deep fusion, and generates a pixel-level cultivated land change detection map through multi-level decoding. The present application can strengthen the expression of spatial homogeneity and time sequence difference, reduce the loss of cross-scale restoration information, and reduce the risk of cultivated land change omission.

The present invention provides a battery management system capable of suppressing misjudgment of a deviation abnormality of a current sensor when a battery current oscillates. This battery management system manages a battery, and is provided with: a main current sensor and a sub current sensor that measure the current of the battery; a first processing unit that calculates a first moving average line on the basis of the plurality of current values measured by the main current sensor, and calculates a first deviation, which is the deviation between the current current value measured by the main current sensor and the first moving average line; a second processing unit that calculates a second moving average line on the basis of the plurality of current values measured by the sub-current sensor, and calculates a second deviation, which is the deviation between the current current value measured by the sub-current sensor and the second moving average line; and a third processing unit that determines, on the basis of the first deviation and the second deviation, a deviation abnormality between the main current sensor and the sub-current sensor.

This invention discloses a method and system for assessing scoliosis based on a depth camera, specifically relating to the field of three-dimensional spinal morphology measurement. The method includes acquiring depth and RGB images of the subject's back, aligning and calibrating them based on camera intrinsic parameters, and outputting aligned depth and RGB images. Using the aligned depth image as input, denoising, cavity filling, and background segmentation are performed to determine the region of interest (ROI) of the spine. The ROI is then back-projected to generate point clouds, which are then voxelized to obtain standardized three-dimensional voxels. By acquiring and aligning back depth and RGB images using a depth camera to generate standardized three-dimensional voxels, a three-dimensional reconstruction network is used to obtain probabilistic voxels of the spine, which are then consistent with the two-dimensional contour to establish reliability. Based on this, the three-dimensional morphology of the spine is reconstructed, and the Cobb angle of the final vertebra is automatically determined, thus outputting reliable quantitative assessment results of scoliosis under radiation-free, low-cost conditions that do not rely on operator experience.

This invention discloses a method and system for acquiring point cloud data and constructing 3D models of karst caves, specifically relating to the field of 3D modeling and edge computing of karst caves. The method includes acquiring continuous point cloud data, pose data, and edge computing status data of a moving vehicle within a natural karst cave; performing temporal alignment and segmented registration on the continuous point cloud data and pose data; outputting a local model and a set of unobserved regions; reading the local model and the set of unobserved regions; performing spatial extension analysis and connectivity deduction on each unobserved region; and outputting a set of candidate structures corresponding to each unobserved region. By generating candidate structures for unobserved regions and calculating disambiguation values, the identification results of key unobserved regions are directly incorporated into subsequent supplementary data acquisition and model update processes, thereby improving the problem of difficulty in confirming key connectivity relationships in natural karst caves.

This invention relates to the field of automotive power battery charge state estimation and energy efficiency monitoring technology, and particularly to a method and system for real-time monitoring and analysis of automotive energy consumption. The method includes: collecting multi-source data from the power battery and vehicle operation to construct an initial monitoring set; generating a synchronization analysis matrix and a working condition switching marker sequence through resampling and anomaly suppression; obtaining a first energy result by performing ampere-hour integration based on the power battery current; extracting polarization characteristic quantities and generating a pseudo-terminal voltage sequence by substituting them into an asymmetric polarization dynamic response model, and combining this with the state estimation equation to obtain a second energy result; calculating the decoupling index and performing weighted calculation to obtain a target energy result; correcting the target energy result and the asymmetric polarization dynamic response model using energy conversion deficit based on energy conservation verification; and performing backtracking correction based on the open-circuit voltage during the resting time interval. This invention solves the problems of instantaneous decoupling and false rebound in energy efficiency data caused by polarization hysteresis under extremely low temperature and drastic changing working conditions.

This invention discloses a method for detecting drying defects in battery electrodes based on infrared thermal imaging, relating to the field of battery electrode technology. The method includes: acquiring a sequence of infrared thermal images of the battery electrode during continuous transport after drying, and simultaneously acquiring the spatial position and motion state information of the image frames to construct a spatial thermal image data array bound to the frame index; performing boundary extraction and contour analysis on the thermal image data array to identify boundary disturbance regions and generate corresponding structural deformation target regions; within the structural deformation target regions, extracting thermal diffusion directionality information based on temperature distribution characteristics in the multi-frame image sequence, and identifying abnormal thermal diffusion regions. This invention identifies local abnormal heat conduction regions by simultaneously extracting boundary disturbance features and temperature change vector information from the thermal image sequence, constructing a main thermal diffusion direction vector field, and calculating a thermal diffusion symmetry breaking factor based on a preset factor fusion rule.

The application discloses a machine learning-based dry-hot valley region soil organic matter and pH spatial distribution prediction method, relates to the technical field of soil spatial information and agricultural resource monitoring, and is based on a pixel and aligns multi-source covariates monthly, constructs machine learning and geostatistics three-line mutual evidence and quantile uncertainty; adopts a bidirectional circulation layer and time sequence self-attention coding, integrates a control layer to multi-head attention, and introduces landform unit, slope direction and altitude band condition coding for fusion and correction; generates a prediction interval with coverage guarantee under the condition of leaving one landform unit and one year, drives directional sampling and suitability confidence linkage; through a closed loop of sampling-retraining-revaluation, converges end error, extrapolation curve and coverage compliance, and outputs a high-precision prediction map, an uncertainty map, a sampling priority and a robust suitability partition.