13results about How to "Avoid blind searches" patented technology

The present application relates to the technical field of short-circuit current calculation, and particularly relates to a flexible DC system short-circuit current calculation method based on a physical information neural network, comprising: setting a model input feature vector, the model input feature vector being used to represent a system operating state before a fault and fault information, and performing data preprocessing on the model input feature vector; constructing a hybrid driving calculation model, the hybrid driving calculation model comprising a physical calculation module and a neural network module, and being coupled based on a preset fusion architecture; performing end-to-end training and optimization on the hybrid driving calculation model by using a preset composite loss function; and performing flexible DC system short-circuit current calculation based on the optimized hybrid driving calculation model, so that the problems of poor precision, speed and convergence, poor interpretability and weak generalization ability in the prior art are solved.

The present application relates to the technical field of optical fiber coupling clamping positioning calibration, and discloses a kind of optical fiber coupling clamping positioning calibration method, optical fiber or lens clamping positioning is calculated with standard calibration piece theoretical coupling target position;With grating calibration plate establishes pixel mapping, calculates spatial displacement deviation;Generation multi-degree-of-freedom adjustment instruction controls five degrees of freedom and monitors coupling efficiency;Adopt random parallel gradient descent algorithm closed loop iteration until coupling efficiency reaches threshold value.The whole process closed loop control from initial clamping to final alignment of the optical fiber coupling clamping positioning calibration method improves coupling efficiency and repeat accuracy, calculates theoretical coupling target position by standard piece turnover measurement, thereby eliminating the system deviation of clamp and optical path, and accurate mapping of pixel and physical size is established using grating calibration plate, image distortion influence is eliminated, visual deviation is converted into spatial displacement deviation after driving adjustment, blind search is avoided.

According to the lost property finding method and device, the storage medium and the computer equipment provided by the invention, when the lost property query instruction is received, blind search of information is avoided by determining the lost property information and accurately positioning the matching data range, and the search efficiency is improved. Meanwhile, the lost article information and the feature vectors in the matching data range are matched in multiple dimensions such as article categories, time and space, appearance features and semantic description, the similarity between the lost article and the lost and found information can be considered more comprehensively, and therefore the matching hit rate is increased. Besides, the feature vectors in the lost property database are divided into public, semi-hidden and completely-hidden information according to user types for multi-level display, so that the requirement of information sharing is met, personal privacy is effectively protected, and the problem of privacy disclosure caused by public display of the details of the lost property is avoided. Finally, the feature vectors with the comprehensive matching scores higher than the preset threshold value are pushed to the query terminal, and the accuracy and reliability of information pushing are ensured.

This invention discloses an intelligent defect identification method and system for photovoltaic modules based on multimodal perception, belonging to the field of photovoltaic module inspection technology. The method includes the following steps: a vehicle platform carrying a multimodal sensor array operates along a preset path, simultaneously acquiring RGB images, infrared thermal images, acoustic signals, and electrical parameters of the photovoltaic modules, and a positioning and navigation module labels the data with spatiotemporal tags; an edge computing module filters the acquired data for validity and uploads it to the software processing layer via a communication link; an image preprocessing module corrects the images, and a multimodal registration module aligns the RGB images and infrared thermal images at the pixel level; an AI inference engine fuses texture features and temperature features, and identifies the fault type based on a preset logical discrimination strategy. This invention provides an intelligent identification system capable of fusing multi-source data to achieve high-precision, in-depth diagnosis and accurate location of photovoltaic module faults.

The invention relates to an express item video positioning method, computer equipment, a computer readable storage medium and a computer program product, which can be applied to the technical field of logistics. The method comprises the following steps: in response to a trigger operation for an abnormal event, determining a camera end identifier corresponding to a target express according to logistics routing information of the target express; according to a shooting time identifier in the unstructured information of the target express, determining a shooting time period corresponding to the shooting end identifier; obtaining candidate image feature vectors from an express feature database of a cloud system according to the camera end identifier and the camera time period; determining a target image feature vector of the target express item from the candidate image feature vectors according to similarity information between the reference image feature vector and the candidate image feature vectors; and positioning a target video of the target express item from the express item video according to a target camera terminal identifier and express item positioning information associated with the target image feature vector. By adopting the method, the express video positioning efficiency can be improved.

This invention discloses a musculoskeletal robot control method for table tennis rallies. The method includes: at the beginning of each stroke cycle, using a trajectory planner based on a physics model to predict the trajectory of the table tennis ball, and calculating the expected hitting position, hitting time, hitting speed, and hitting direction of the racket held by the musculoskeletal robot based on a preset hitting plane and target landing point, forming a high-level control command; inputting the high-level control command and the current environmental observation state into a pre-trained reinforcement learning policy network, outputting the target joint position of the musculoskeletal robot; and using an adaptive controller based on forward kinematics to map the target joint position to the activation signals of each muscle, driving the musculoskeletal robot to perform actions. This invention solves the control problem of high-dimensional musculoskeletal systems in complex dynamic tasks through a hierarchical architecture, achieving continuous table tennis rallies with high success rate and low muscle metabolic cost.

This invention relates to a method and apparatus for tuning control parameters of magnetic levitation bearings, a storage medium, and an electronic device, belonging to the technical field of magnetic levitation control parameters. The tuning method includes: S1, constructing a mathematical model of a PID parameter optimization problem; S2, generating a Bayesian optimization sample set, fitting a Gaussian process surrogate model, and globally exploring to generate a PID parameter subspace; S3, executing a dung beetle optimization algorithm within the subspace to obtain the optimal point (x) within the subspace. db y db S4, add the optimal point to the Bayesian optimization sample set and update the Gaussian process model; S5, repeat steps S2 to S4 until the condition is met, and output the optimal PID parameter x. db * By combining the global exploration capability of Bayesian algorithms with the local development capability of the dung beetle algorithm, the Bayesian optimization reduces the dependence on the initial sample size, while the dung beetle algorithm accelerates convergence within the subspace, thus balancing efficiency and accuracy.

The invention is suitable for the technical field of computer vision, and provides a target detection model training method, a target detection method and electronic equipment, the target detection model comprises a prior network, an encoder and a decoder, and the method comprises the following steps: obtaining a training image; the training image is input to the prior network, prior information output by the prior network is obtained, and the prior network is used for detecting the position of a target in the training image; inputting the training image into the encoder for feature extraction to obtain image features output by the encoder; inputting the prior information and the image features to the decoder for decoding to obtain a first detection result, the first detection result being used for indicating the position and category of a target in the training image; and updating the target detection model based on the first detection result to obtain a trained target detection model. According to the invention, the robustness and generalization of the target detection model can be improved.

This utility model discloses an integrated emergency lighting device capable of real-time detection of human body location. The device includes an emergency lighting fixture body, which integrates a millimeter-wave radar sensing module, a thermal imaging sensor, a voice recognition module, a CO gas sensor, a PLC control module, a drive unit, and an audible alarm. The millimeter-wave radar sensing module, thermal imaging sensor, voice recognition module, CO gas sensor, drive unit, and audible alarm are all electrically connected to the PLC control module. The emergency lighting fixture body has two evacuation lights, each electrically connected to the PLC control module via a drive unit, forming independent working circuits that do not interfere with each other. This utility model can promptly monitor and locate the position of trapped personnel, providing firefighters with accurate rescue information, improving rescue precision, avoiding blind searches, shortening rescue time, and reducing casualties and property damage.

The invention belongs to the technical field of primary frequency modulation, and provides an energy storage primary frequency modulation method and system based on model-data hybrid driving, and the method comprises the steps: building an equivalent mathematical model in which an energy storage battery system participates in primary frequency modulation; carrying out online identification on the static parameters by adopting a weighted recursive least square method; constructing a second-order RC equivalent circuit model of the energy storage battery and carrying out SOC preliminary estimation; inputting the SOC preliminary estimation result and the battery operation data into a pre-trained deep learning model to carry out refined SOC estimation; and based on a refined SOC estimation result, establishing an SOC amplitude limiting mechanism by using a Logistic function, and applying an energy storage output constraint. According to the invention, energy storage output amplitude limiting under different frequency modulation requirements can be obtained according to the SOC.

The invention relates to the technical field of submarine cable inspection, in particular to a submarine cable inspection method and system based on an underwater inspection robot. According to the method, the physical characteristics of the metal properties of the cable are used as the fusion key, the corresponding visual image data pairs can be accurately found through consistency verification of the physical properties even if a plurality of interferents with similar geometrical shapes exist or errors exist in pose calculation, and the fusion reliability is remarkably improved. Meanwhile, the use of the acoustic trusted area mask provided by the embodiment provides powerful prior information for the search and verification of the visual image, avoids blind search in massive visual image data, and improves the efficiency of data fusion.

The present application belongs to the technical field of quantum computing and artificial intelligence, and particularly relates to a quantum neural network architecture search optimization method and system and application thereof in agricultural image classification. The present application adopts a quantum coarse screening-classical fine tuning method, efficiently excludes architectures with poor performance or not meeting resource constraints in the search space through Grover quantum pre-screening, quickly narrows the search funnel to the potential high performance area, avoids blind search of the classical evolutionary algorithm in a huge space, and greatly shortens the overall search period. The present application can finely optimize multiple targets such as precision, delay and resource consumption, and the obtained model architecture can effectively balance the problem of "global exploration" and "local exploitation" in the search process, so that the model architecture has the characteristics of high precision and low delay; meanwhile, the preset value lightweight constraint ensures the lightweight and high performance of the architecture, has good compatibility, and can completely compatible with the existing agricultural IT infrastructure and edge devices.

The application relates to the technical field of picture processing comparison, and discloses a feature image automatic detection comparison method, which comprises the following steps: acquiring an object image containing object features, performing feature detection and segmentation on the object image by using a feature detection algorithm formulated according to the characteristics of the object features, obtaining source feature data and storing the source feature data; acquiring a to-be-verified image containing a to-be-verified object, matching corresponding source feature images and source feature data according to the identification information of the to-be-verified object; inputting the to-be-verified image into a feature detection algorithm corresponding to the source feature data to obtain to-be-verified feature data; performing feature comparison based on the source feature data and the to-be-verified feature data by using a comparison algorithm to output a comparison result; wherein the feature detection algorithm comprises cross-scale feature extraction and a dynamic feature weight strategy, can comprehensively capture the detailed information and overall distribution law of the object features, reduces the interference of interference factors on feature extraction and comparison, and improves the comparison accuracy and efficiency.