27results about How to "Realize precise identification" patented technology

The invention provides a salt rock expansion point identification method and electronic equipment, and relates to the technical field of geological engineering monitoring. The method comprises the following steps: carrying out sliding sampling on acoustic emission data in an initial search interval according to a first preset window length to obtain a plurality of initial window data segments; determining an initial expansion point position according to the plurality of initial window data segments; determining a target search interval according to the initial capacity expansion point position; wherein the target search interval is smaller than the initial search interval; performing sliding sampling on the acoustic emission data in the target search interval with a second preset window length to obtain a plurality of target window data segments; wherein the second preset window length is smaller than the first preset window length; and determining a target expansion point position according to the plurality of target window data segments. Acoustic emission data is adopted as a data source and can be used for real-time monitoring; and meanwhile, a two-section identification method of rough search and fine search is adopted, so that accurate identification of the expansion point can be realized.

The invention discloses a facial expression recognition method and device based on multimode attitude measurement learning, and a storage medium, relates to the technical field of facial expression recognition, and solves the technical problems of weak generalization ability, insufficient cross-modal feature fusion and easy forgetting of pre-training knowledge due to data scarcity in the existing facial expression recognition technology. The method comprises the following steps: acquiring image data containing facial expressions, preprocessing the image data, constructing a text prompt word set, and constructing triple training samples of anchor points, positive samples and negative samples; a visual cue word vector is designed to be injected into a Transform layer of a pre-trained CLIP-ViT-B / 32 model, and model trunk parameters are frozen to only train visual prompt and LayerNorm layer parameters; triple loss and classification loss combined optimization is adopted; in the reasoning process, the expression category is determined by calculating the similarity between the to-be-recognized image features and various expression text prompt features, pre-training semantic knowledge and cross-modal adaptation ability of CLIP are fully utilized, and the expression recognition precision, robustness and zero sample generalization ability are remarkably improved.

This invention provides a method and system for constructing a 3D model of a construction site by combining multiple data sources. It relates to the field of data processing technology. The method includes: retrieving satellite imagery at a preset resolution to obtain a full-area stereo image pair dataset; performing radiometric correction to generate a regional surface model and a regional elevation model; performing raster-level modeling defect detection to locate distributed blind zone blocks; fitting obstacle avoidance paths to generate optimized obstacle avoidance flight trajectories; driving a UAV to perform multi-angle surround scanning to directionally collect dense point clouds in the distributed blind zones; constructing a distributed blind zone model; and performing microscopic overlay to output a 3D model of the construction site. This invention solves the technical problems of existing construction site modeling methods, which have limitations in coverage, modeling efficiency, and data real-time performance, and rely on a single data source, failing to provide a comprehensive and accurate 3D model of the construction site, thus affecting construction progress and decision-making.

This invention relates to the field of image data processing technology, specifically to a deep learning-based intelligent identification system and method for blasted rock mass size. The method includes: acquiring images of rock accumulation at a blasting site using an image acquisition device; importing these images into a deep learning-based intelligent identification system for blasted rock mass size; analyzing the blasted rock mass size data; generating blasted rock mass size curves; and performing a reasonable analysis of the blasting effect. The intelligent identification system for blasted rock mass size includes an image preprocessing module, a rock mass size analysis module, and a result visualization module. The image preprocessing module performs image enhancement, noise reduction, and segmentation; after feature extraction via a convolutional neural network, the rock mass size identification module performs rock mass segmentation, classification, and size measurement; the rock mass size analysis module statistically generates distribution curves; and the visualization module presents the results. This identification system achieves automatic, rapid, and accurate identification of rock mass size in complex scenes, effectively improving the efficiency and accuracy of blasting effect evaluation.

The application provides a face recognition method and device in adaptive scenes, a storage medium and an electronic device. The method comprises: obtaining a to-be-recognized face image; determining that the current light scene is a backlight scene when the image brightness value of the to-be-recognized face image is greater than a brightness threshold; determining that the current light scene is a dark light scene when the image brightness value of the to-be-recognized face image is less than or equal to the brightness threshold and the image gain value is greater than an upper limit of the gain; and determining that the current light scene is a normal light scene when the image brightness value of the to-be-recognized face image is less than or equal to the brightness threshold and the image gain value is less than or equal to the upper limit of the gain. The image brightness value and the image gain value of the to-be-recognized face image in each light scene are adaptively compensated for the backlight scene, the dark light scene and the normal light scene respectively. The present scheme realizes accurate recognition of the to-be-recognized face image obtained in different light scenes.

The application discloses a mud generator torque control method and system, and belongs to the technical field of well drilling engineering. The method comprises the following steps: performing phase identification on three-phase alternating current output by a mud generator through a digital processor, generating a PWM driving signal based on phase information, controlling the conduction and shutdown of a power switch tube connected with a generator winding, adjusting the current of a generator stator winding, and completing the control of the output torque of the generator. The application performs phase identification on three-phase alternating current through a digital processor and generates a PWM driving signal, realizes accurate control of the output torque of the generator, directly adjusts the current of the stator winding, and avoids the heat dissipation and space problems caused by the traditional resistance dissipation mode.

The application discloses a bridge service linear identification method based on multiple three-dimensional laser scanning, comprising the following steps: defining and acquiring the bridge girder linear; acquiring the three-dimensional point cloud data of the bridge under the condition of non-interrupted traffic in different service periods; converting the three-dimensional point cloud data of the bridge obtained through multiple scanning under the same coordinate system; obtaining the bridge service linear with statistical significance and meeting the preset confidence condition through regression prediction; comparing the bridge service linear in different periods to identify the spatial form change of the bridge girder, evaluate the bridge structure state in the bridge service period, and track the spatial form change of the bridge in the bridge service period. The application realizes the accurate identification of the bridge service linear, solves the problem of insufficient spatial information acquisition in the traditional measurement method, and realizes the bridge service linear identification with high confidence in view of the complex vibration noise problem of the point cloud array caused by the bridge vibration, thereby improving the accuracy of the bridge service state control.

The present application relates to the field of particle detection, and specifically provides a method and device for identifying and counting ultrafine particles based on photoelectric effect, and a storage medium, comprising: obtaining photoelectric signals generated by a flow of particles to be detected passing through a photoelectric detection array; extracting original features of the photoelectric signals, combining time domain decomposition and frequency domain transformation to construct an initial signal feature matrix; performing multi-scale tensor decomposition and signal component screening on the initial signal feature matrix to generate enhanced photoelectric signals; determining a polarization coherence angle interval based on the enhanced photoelectric signals, extracting polarization energy features, analyzing polarization state conversion characteristics, and establishing a three-dimensional polarization feature fingerprint; performing adaptive segmentation processing on the enhanced photoelectric signals, extracting oscillation feature spectrum, and combining the three-dimensional polarization feature fingerprint to obtain a feature correlation matrix; and constructing a multi-dimensional classification criterion based on the feature correlation matrix, and determining a final particle count result accordingly. The present application improves the accuracy and recognition ability of ultrafine particle counting, and is particularly suitable for high-density micro-nano particle detection in complex environments.

The invention discloses a dynamic pressure balance regulation and control method and system for a water conveying pipeline, and relates to the technical field of fluid conveying control, and the method comprises the following steps: collecting synchronous pressure data of a plurality of monitoring points arranged along the water conveying pipeline, endowing a unified time scale, constructing a pressure change sequence sorted according to time, and a pressure sudden change time point is marked in the pressure change sequence. By constructing a pressure change sequence and a propagation delay record with a unified time scale, the method achieves the recognition of a pressure wave propagation direction and an echo time point, and prevents a reflection echo from being misjudged and repeatedly triggered and adjusted. Meanwhile, by generating an adjustment time window and implementing reverse fine adjustment and asymmetric rollback control, the adjustment action and echo superposition process is interrupted, pressure oscillation is restrained, and the operation stability and safety of the water conveying pipeline are improved.

The application discloses a kind of underwater vehicle control system competition failure reliability evaluation method, belong to underwater vehicle reliability evaluation technical field, for underwater vehicle control system competition failure reliability evaluation, the multiple-source monitoring of underwater vehicle control system is preprocessed, and the reliability model of underwater vehicle control system is constructed;Non-time-lag parameter and time-lag related parameter are constructed, and the estimated value of non-time-lag parameter and time-lag related parameter is calculated using least square method, statistical inference and time-lag structure identification method;Introduce alternate iteration optimization mechanism to carry out parameter estimation, and the reliability of underwater vehicle control system is calculated.The application realizes the accurate modeling and reliability evaluation of the competition failure degradation process and impact coupling effect of underwater vehicle control system by constructing dynamic self-adaptive feature extraction model and time-lag related degradation theory, and using parameter estimation and alternate iteration optimization strategy.

The present application relates to the technical field of intelligent power distribution equipment and power distribution automation, and particularly relates to an automatic topology identification method of a modular power distribution box; the process comprises: collecting connection behaviors of module interface insertion, mechanical locking and main circuit conduction state to form a connection behavior sequence; generating a structure-derived connection behavior sequence based on the power supply path of the module in the box and the correlation module relationship; checking the time sequence correspondence relationship and state closure relationship between the interface level and the structure-derived level sequence through cross-level consistency rules to form a topology relationship determination result; performing local topology update according to the determination result to update the topology structure mapping of the modular power distribution box; the present application overcomes the recognition dead zone and electromagnetic coupling interference caused by low impedance and high collinear branch, can accurately determine the real physical connection relationship between the modules, supports plug-and-play of the modules and dynamic adjustment of the local topology, and improves the system topology identification precision and operation and maintenance management efficiency.

The invention provides a DCS-based intelligent start-stop control method and system for a photo-thermal power generation system, and relates to the technical field of photo-thermal power generation, and the method comprises the steps: collecting real-time operation data through the DCS, extracting features through data standardization and wavelet transformation, extracting time sequence features through a variational recurrent neural network, generating a reconstruction sequence through an auto-encoder, and screening key time slices. And extracting decision rules by using knowledge distillation, constructing a strategy network, generating a control strategy, and deploying the control strategy to a DCS control unit. Start-stop control precision of the photo-thermal power generation system can be improved, system operation risks are reduced, and the service life of equipment is prolonged.

This invention discloses an automated end-to-end management system for invoices with anti-counterfeiting tracking capabilities, including a traceability anchor point generation module, an anti-counterfeiting verification and classification module, a watermark injection and circulation module, and a cross-system verification module. This invention generates physical feature fingerprints by fusing hash values ​​and QR code fault-tolerant distribution feature vectors, binding these fingerprints to invoice data to form traceability anchor points; achieves high-precision format verification based on pixel density feature comparison technology using micro-grid arrays; utilizes multi-dimensional semantic tags for intelligent classification and storage; employs invisible circulation node watermark technology to ensure cross-system circulation security; and combines a traceability graph reconstruction mechanism to achieve accurate tampering accountability. This invention effectively improves the reliability of anti-counterfeiting capabilities, achieves end-to-end tracking and real-time deduplication, and provides secure and efficient technical support for enterprise financial and tax management.

The invention provides a self-adaptive blade deicing system and method based on multi-modal perception and AI decision, and is applied to the technical field of wind power generation, the system comprises an acquisition device used for acquiring blade state data; the mode selection module is used for adopting different operation modes according to the external environment state data, correspondingly starting acquisition devices in different areas on the blade, and acquiring real-time blade state data; the data processing and communication module is used for performing primary feature extraction on the real-time blade state data to obtain primary feature data and synchronously transmitting the primary feature data to the AI decision module; the AI decision module is used for generating a deicing measure of the blade based on the primary characteristic data, performing ice melting process prediction to obtain a secondary icing risk, and generating an enhancement strategy in combination with the secondary icing risk; and the partition execution module is used for performing partition deicing according to the enhancement strategy. The problems that a traditional deicing system is high in energy consumption, fuzzy in perception and prone to causing secondary icing are effectively solved.

The application provides a method and device for generating a wiring diagram based on CAD image recognition, comprising the following steps: image preprocessing; electrical element recognition in the image; text label recognition in the image; line segment recognition in the image; according to the blocking characteristics of the power grid control image, the obtained elements, texts and connecting lines are locally clustered and combined, the combined image is aligned and spliced, and a correlation is constructed; and the recognized electrical elements, texts and correlation are written into a G file to generate a wiring diagram. The user can conveniently and efficiently realize power grid control diagram drawing automation by using the method, effectively improve work efficiency, and meet the overall goal of upgrading and transformation of traditional power grid dispatching business to automation, intelligence and efficiency.

This invention discloses a cable condition identification and early warning method based on a quadruped robot for multi-sensor fusion inspection. The invention utilizes a quadruped robot to scan the tunnel environment in real time, navigating along a preset inspection route. Visible light and infrared thermal images are simultaneously acquired using a dual-spectrum gimbal, and pixel-level alignment is achieved after image enhancement and ORB feature registration. The visible light image is segmented using a U-Net network, and the cable region temperature matrix is ​​extracted from the infrared thermal image, with the average temperature calculated. Anomaly detection is performed based on multi-level temperature criteria. When an abnormal temperature is detected, a partial discharge sensor mounted on a six-axis robotic arm is activated for fixed-point re-testing, detecting discharge amplitude and frequency for multi-level early warning. A dynamic threshold update mechanism is introduced to adjust the judgment threshold. This invention accurately identifies defects such as cable overheating and insulation degradation, providing timely early warning of potential faults and effectively ensuring the stable operation of high-voltage cable lines.

The invention relates to the technical field of rail transit train operation control and navigation positioning, and discloses a seamless integrated positioning system and method in a complex railway tunnel environment, and the system employs a three-dimensional laser radar to construct a local environment map in real time, detects the physical shielding condition of an ultra-wideband signal propagation path through a light tracing technology, and achieves the real-time positioning of the ultra-wideband signal. And the non-line-of-sight propagation state is accurately identified. And based on an identification result, the system dynamically adjusts a measurement noise covariance, reduces the weight of non-line-of-sight measurement under a factor graph optimization framework, and fuses IMU pre-integration and LiDAR odometer data to carry out real-time pose calculation. In addition, high-precision satellite positioning states of the entrance and exit of the tunnel are used as double-end boundary constraints, and global adjustment is carried out on the whole-process track of the tunnel. According to the method, the non-line-of-sight error is effectively inhibited from the physical level, dead reckoning accumulated drifting is eliminated, and the positioning robustness and precision of the train in the complex tunnel environment are improved.

The invention discloses a video monitoring scene intelligent analysis and event classification system based on semantic segmentation. The system comprises a data preprocessing module, a preliminary classification module, a weight generation module and a correction and detection module. The data preprocessing module is used for acquiring and preprocessing original video data of a video monitoring scene, and extracting scene semantic features of the preprocessed video data through a semantic segmentation model; and the preliminary classification module is used for inputting the scene semantic features into a classifier for performing preliminary event classification to obtain an initial classification result. The invention relates to the technical field of video monitoring and intelligent analysis. According to the video monitoring scene intelligent analysis and event classification system based on semantic segmentation, the coupling degree between modules of the system is low, the semantic category number, the classifier structure and the updating strategy can be freely configured according to different monitoring tasks, and the system is suitable for various application scenes such as traffic flow analysis, personnel gathering monitoring and industrial operation anomaly detection.

The application discloses a self-powered pressure sensor and a preparation method thereof, and the self-powered pressure sensor comprises a zinc electrode, a copper electrode, an ionic gel electrolyte film and a packaging layer provided with a gas escape gap, the upper surface of the ionic gel electrolyte film is provided with a hemispherical microstructure array, and the lower surface is provided with a microgroove or micro-pit array. The self-powered pressure sensor provided by the application solves the problems of low output performance and insufficient cycle durability caused by the deep coupling of energy supply and sensing function in traditional electrochemical self-powered sensors by means of the ionic gel electrolyte film with double-layer interlocking microstructures, high conductivity and good mechanical properties, and realizes the balance between high-sensitivity pressure sensing and high-power energy output. The self-powered pressure monitoring system constructed on the basis of the sensor can work independently without an external power supply, and can realize accurate identification of human motion rehabilitation actions in combination with a convolutional neural network model.

The invention discloses a snow melting and temperature rising process disaster chain hidden danger recognition method based on big data. The method comprises the following steps that remote sensing snow covers, radar wet snow lines, daily temperature, flow and ground surface deformation data of a research area are obtained and preprocessed; an improved Bayesian online change point detection algorithm is adopted to determine a real-time node triggered by snow melting temperature rise; calculating a snow cover area change rate and a wet snow line change rate to identify a snow melting sensitive area; analyzing the surface deformation data to determine a surface acceleration sensitive area; analyzing the flow data to determine a runoff mutation sensitive area; constructing a disaster chain hidden danger transmission space-time structure model of snow melting triggering, surface acceleration and runoff response; and identifying a disaster chain hidden danger level in real time based on a space-time structure model. According to the invention, real-time and quantitative recognition of regional scale disaster chain hidden dangers is realized, and the accuracy and timeliness of disaster early warning are improved.

This invention relates to the field of real-time resource scheduling technology, specifically a real-time scheduling method for lane control robots based on multi-dimensional sensor data. The method includes: dividing management areas for arrival and drop-off robots; acquiring multimodal sensor data from the management areas, extracting features, and fusing them using a self-attention mechanism to obtain a fused feature vector; obtaining a drop-off scenario complexity score and a departure intention score through a drop-off scenario evaluation model; simultaneously calculating the real-time resource occupancy index for each lane; calculating differentiated allowed dwell time based on the drop-off scenario complexity score and the overall resource occupancy status of the drop-off platform; and providing early warnings and management based on the drop-off scenario complexity score, departure intention score, real-time resource occupancy index, differentiated allowed dwell time, and empty space detection results. This invention improves flexible decision-making capabilities and resource utilization, and is suitable for real-time resource allocation and decision support in complex traffic scenarios.

This invention discloses a method for modeling the propagation characteristics of electromagnetic interference (EMI) in integrated primary and secondary distribution switches, belonging to the field of electromagnetic interference prediction and simulation technology. The method includes: obtaining a port response characteristic matrix through three-dimensional electromagnetic modeling and field-circuit co-simulation based on the electromagnetic coupling mechanism of a voltage sensor under high-frequency conducted interference; obtaining a disturbed output characteristic matrix of a current sensor through joint magnetic field-circuit modeling of a Rogowski coil based on the coupling path and frequency response characteristics in the port response characteristic matrix; and obtaining a system-level EMI propagation network matrix through field-circuit coupling modeling of the cable spatial magnetic field and multi-port crosstalk using the port response of the disturbed output characteristic matrix of the current sensor. Finally, a multi-physics integrated model of the entire integrated primary and secondary switch is obtained based on the coupling coefficient and propagation path in the system-level EMI propagation network matrix, resulting in a full-path simulation model of EMI propagation. This invention solves the problem that existing technologies cannot accurately model the EMI propagation characteristics of integrated primary and secondary distribution switches.

The present application relates to a kind of particle identification and detection method based on deep learning improved algorithm, it includes the following steps: using electric sensitive area method obtains the electric signal of multiple particle;Extract the time domain, frequency domain and time-frequency domain characteristic parameters of electric signal;Characteristic parameters are combined to enhance the difference between characteristics;Screening characteristic parameters to reduce the amount of operation;Neural network and recurrent neural network model are weighted fusion;Using fusion model to test multiple repeat for the particle to be measured, using voting mechanism to determine the final classification result;The concentration, quantity proportion and particle size distribution of each kind of particle are output.The particle identification and detection method provided by the present application can realize the measurement of particle type, concentration and particle size distribution in water, improve the accuracy, accuracy and stability of water quality monitoring, provide more comprehensive information for water plant operation and environmental monitoring, so that the particle identification and detection method can be better applied to practical application.

The embodiment of the application provides a dynamic resource scheduling method and device based on an AI agent, accurate demand identification is realized through feature quantization and state vector construction. A scheduling mechanism is constructed, combined with load prediction and strategy formulation, to establish a reliable resource allocation strategy. Feedback optimization is introduced, through state monitoring and reinforcement learning, to ensure continuous improvement of scheduling. This method effectively solves the shortcomings of traditional technology in demand identification, resource allocation and feedback optimization, and provides technical support for dynamic resource scheduling.

This invention provides a physical simulation experimental device and method for fine exploration of water-bearing sections in underground rock strata. The device includes a water storage tank connected to a water pump. The water pump is connected to one end of multiple water supply valves via a one-way valve. Each water supply valve has a corresponding rock chamber at its other end. These rock chambers are vertically mounted on the sidewall of a simulated borehole pipe. Flange end caps are installed at both ends of the simulated borehole pipe. A return water inlet sealing head is installed on the flange end cap at the tail end of the simulated borehole pipe. The return water inlet sealing head is connected to the water storage tank via a return water valve. An energy storage and voltage regulator is also installed between the water storage tank and one end of the multiple water supply valves. An armored optical fiber cable for acoustic wave sensing is coaxially installed inside the simulated borehole pipe. This invention can effectively simulate various geological and hydrological conditions in actual underground rock strata environments. By analyzing the variation patterns of acoustic signals, it improves the accuracy of exploration of water-bearing sections, providing reliable technical support for coal mine water hazard prevention and safe production.

The embodiment of the invention provides a dynamic resource scheduling method and device based on an AI agent, and the method and device achieve the precise recognition of demands through feature quantization and state vector construction. And a scheduling mechanism is constructed, and a reliable resource allocation strategy is established by combining load prediction and strategy formulation. And feedback optimization is introduced, and continuous improvement of scheduling is ensured through state monitoring and reinforcement learning. According to the method, the defects of the traditional technology in the aspects of demand identification, resource allocation, feedback optimization and the like are effectively overcome, and technical guarantee is provided for dynamic resource scheduling.

The application discloses a self-powered pressure sensor and a preparation method thereof, and the self-powered pressure sensor comprises a zinc electrode, a copper electrode, an ionic gel electrolyte film and a packaging layer provided with a gas escape gap, the upper surface of the ionic gel electrolyte film is provided with a hemispherical microstructure array, and the lower surface is provided with a microgroove or micro-pit array. The self-powered pressure sensor provided by the application solves the problems of low output performance and insufficient cycle durability caused by the deep coupling of energy supply and sensing function in traditional electrochemical self-powered sensors by means of the ionic gel electrolyte film with double-layer interlocking microstructures, high conductivity and good mechanical properties, and realizes the balance between high-sensitivity pressure sensing and high-power energy output. The self-powered pressure monitoring system constructed on the basis of the sensor can work independently without an external power supply, and can realize accurate identification of human motion rehabilitation actions in combination with a convolutional neural network model.