101 results about "Algorithms performance" patented technology

The invention provides a robot vision object semantic understanding and posture generation method based on a large model, and belongs to the technical field of image processing. Comprising the following steps: S1, inputting an image and a 3D model; s2, object detection; s3, multi-modal feature alignment is carried out; s4, diffusion model sampling; s5, performing geometric screening; s6, carrying out NeRF (New Random Field) morphological modeling; s7, optimizing the optical flow; s8, joint loss calculation; s9, confidence coefficient analysis; and S10, outputting the attitude junction. According to the invention, through an innovative single-view rendering-optical flow optimization closed loop strategy, the calculation overhead is significantly reduced and the estimation precision is improved. According to the method, the algorithm performance in a complex scene is remarkably improved through multi-scale feature fusion and a confidence decomposition strategy.

The invention discloses a mechanical arm simulation system supporting industrial quality inspection, which relates to the technical field of industrial automation and comprises a multi-physics field simulation engine module, a digital twin interface module, a defect sample library module and a cooperative control module. According to the method, traditional entity mechanical arm debugging is replaced by rigid body dynamics simulation based on a physical engine, loss and production line shutdown caused by repeated trial and error of equipment are avoided, rapid construction of the digital twin environment of the production line is achieved by means of an efficient CAD model conversion process and a lightweight loading technology, and high-fidelity optical simulation is achieved through a parameterized defect model library and high-fidelity optical simulation. Various common industrial defects and complex optical effects are covered, the modeling bottleneck of a quality inspection scene is broken through, real-time closed-loop interaction of visual inspection and motion control is realized by using thread separation and an efficient communication technology, and the recognition robustness of a visual algorithm to tiny defects is guaranteed; the reliability of algorithm verification is improved through the quantitative output of algorithm performance indexes and the comprehensive test of extreme working conditions by the simulation system.

The invention belongs to the technical field of traditional Chinese medicine information processing, and provides a traditional Chinese medicine disease prescription data base construction method and system.The method comprises the steps that five types of data including ancient literatures, modern medical records, prescription data, traditional Chinese medicine resources and medical record manuscript images are accessed through an extensible adapter frame; a traditional Chinese medicine term service platform is constructed by using multi-source term resources, and standardized data is generated from the analyzed data; the method comprises the following steps: extracting diseases, syndromes, prescriptions and traditional Chinese medicine entities by adopting a rule engine and a domain self-adaptive entity recognition model, on the basis of a public traditional Chinese medicine standard knowledge graph, aggregating neighbor node information through GAT, and introducing a contrast learning optimization entity relationship classifier to obtain an enhanced entity vector and classify a semantic relationship; constructing a knowledge graph based on the extracted semantic relationship; and carrying out data quality and algorithm performance monitoring and tracing on data access analysis, standardization processing, knowledge extraction and graph construction. Uniform access and standardized integration of multi-source heterogeneous traditional Chinese medicine data are realized.

The application discloses an image conversion method and device, a computer storage medium and a program product, and relates to the technical field of semiconductors. The method comprises the following steps: obtaining a target binary image, the target binary image comprising N closed graphics; according to binary assignment information of pixel points in the target binary image, each target pixel point located on a graphic edge contour of a target closed graphic is determined by sequentially traversing the pixel points in the target binary image; based on each target pixel point, edge contour coordinates corresponding to the target closed graphic are extracted until edge contour coordinates of the N closed graphics are obtained; and the target binary image is converted into corresponding target artwork based on the edge contour coordinates of the N closed graphics. According to the embodiment of the application, the image conversion can be more efficient and accurate, and the overall image conversion efficiency, algorithm performance and accuracy are effectively improved.

The invention relates to a method for realizing machining element guide treatment in plate rotating and turning machining, which comprises the following steps that three directions and two edge distances are defined for each machining element, the three directions comprise a machining surface, a reference surface and a direction surface, and the two edge distances comprise a longitudinal edge distance and a transverse edge distance; in the plate rotating or turning process, direction changing is conducted on the machining face, the reference face and the direction face according to a rotating algorithm; and determining the relative position of the processing element on the plate according to the transformed direction surface and the corresponding distance. By the adoption of the method and device for achieving machining element guide processing through rotating and turning plate machining of the plate, the processor and the readable storage medium of the processor, under the condition that shape information of the elements is not changed, the plate is rotated, quadrant switching is conducted, data consistency is guaranteed, error transformation caused by data change is reduced, and the machining efficiency is improved. The rotation algorithm performance is greatly optimized, and logic processing for obtaining the machining tool path according to the element information is simplified.

The application discloses a kind of high-precision Beidou time service master clock devices based on double-chip atomic clock fusion, it relates to timing synchronization technical field, the device includes satellite receiving module, master control processing module, double-chip atomic clock module, clock taming module, time-frequency output module, display control module and power module;Each module is cooperated by work mode automatic switching mechanism Work;Double-chip atomic clock module adopts two pieces chip level atomic clock and constitutes master-slave redundant architecture, and the phase and frequency deviation information of main atomic clock and slave atomic clock are fused and mutually checked by extended Kalman filtering;The taming algorithm in clock taming module is formed by extended Kalman filtering link, PID controller link, digital phase-locked loop link cascade;Work mode automatic switching mechanism is according to the automatic switching of work mode according to the operating state of the device;Overcome the insufficient of time service precision, low reliability, taming algorithm performance is limited and so on in prior art.

The application discloses a quantum image multiplication operation method and device and electronic equipment, and relates to the field of quantum operation, and solves the problem of low performance of a quantum image multiplication operation algorithm in the prior art, and the technical solution points are as follows: a plurality of quantum images with shared position information and in a superposition state are prepared based on a NEQR quantum image representation model; any two quantum images in the plurality of quantum images are subjected to multiplication operation processing based on a quantum multiplier, and a matrix of pixel values of the quantum image is obtained; information of each pixel point in the matrix of pixel values of the quantum image is extracted, and the information is converted into classical image information; wherein, the operation steps of quantum multiplication are improved according to binary multiplication, the quantum multiplication improved based on binary multiplication is designed by using a Toffoli gate and a quantum full adder, and a quantum multiplier is obtained. The application improves the performance of a quantum image expression algorithm, and makes it easier to simulate under a classical computer.

This invention discloses a deep learning-based method for detecting aircraft skin defects, belonging to the field of defect detection technology. The method includes: acquiring a target image; the target image being an image of an aircraft skin defect; inputting the target image into a defect detection model to obtain an image with defect boundaries; the defect detection model is trained using an augmented dataset based on Transformer and convolutional neural networks; the augmented dataset is determined based on an aircraft skin defect dataset and a brightness generation branch network. This invention overcomes the dual challenges of extreme lighting variations and significant target scale differences in complex apron environments by introducing a collaborative mechanism between the brightness generation branch and the multi-scale detection branch. This significantly improves the model's accuracy and recall for detecting aircraft skin defects, reducing the false negative rate while maintaining high inference speed, thus enhancing the performance of the detection algorithm.

The present application relates to the technical field of tire mechanics, in particular to a multi-strategy improved tire model parameter identification method and system considering envelope constraints, a multi-strategy tire model parameter identification method considering envelope constraints is fused with tire data preprocessing, covers pure longitudinal working conditions, pure lateral working conditions, pure lateral return-to-normal working conditions, combined longitudinal working conditions and combined lateral working conditions, and relates to all working conditions of tire testing; the model parameter method is determined in a smaller parameter range, which can greatly reduce the search range and improve the convergence accuracy and speed; and the original AO algorithm is improved in a targeted multi-optimization strategy to improve the algorithm performance.

The application discloses a peanut-shaped magnetorheological capsule robot living tissue sampling controller based on IGJO, which comprises mechanical analysis of a living tissue sampling module for the peanut-shaped magnetorheological capsule robot, research on the driving effect of an external driving permanent magnet on the capsule robot, establishment of the living tissue sampling controller of the capsule robot, design of an improved golden jackal optimization algorithm, introduction of Fuch mapping theory and an elite population strategy in the population initialization stage, optimization initialization of the golden jackal population, introduction of an adaptive inertia weight omega and a searcher position updating strategy in the improved sparrow algorithm in the iterative output stage of the population, repeated optimization of the output solution of the golden jackal population, and effective improvement of the global search capability of the algorithm. The application improves the algorithm performance of the golden jackal algorithm, avoids falling into local optimization, realizes the optimization and adjustment of the living tissue sampling controller parameters of the capsule robot, optimizes the control effect, and improves the control precision.

The invention discloses an industrial software service combination method based on an improved multi-objective grey wolf algorithm, and the method comprises the following steps: building an industrial software resource service combination optimization problem framework, and completing the evaluation of QoS indexes including service time, service cost, service reputation, service delivery quality and the like and availability indexes; an SCOS model with optimal service quality (QoS) and availability as targets is constructed; tent chaotic mapping and a Levy flight improved differential evolution strategy are fused into a multi-target grey wolf algorithm model, the improved multi-target grey wolf algorithm model is used for solving an SCOS model, and the effectiveness and feasibility of the method are verified through algorithm performance verification and example testing. According to the improved multi-target grey wolf algorithm model, the convergence speed can be effectively increased, manual combination is liberated, and the problem of local optimization of industrial software combination is effectively solved.

The invention discloses a multi-mode three-dimensional target detection system and method based on 3D Gaussian sputtering. The system comprises a bimodal feature extraction module, a vector-matrix Gaussian representation module and a graph query detection network. The method comprises the following steps: firstly, inputting a LiDAR point cloud projection and an RGB image to a bimodal feature extraction module, and outputting a feature matrix group; transmitting the generated feature matrix to a vector-matrix Gaussian representation module, decomposing the feature matrix into a vector and a matrix, and generating a 3D Gaussian field through VMGS; transmitting the generated 3D Gaussian field to a graph query detection network, analyzing the Gaussian field, and outputting a target category probability and a 3D bounding box parameter; and finally, recording detection related data on a target detection result display, and evaluating the performance of a detection algorithm. According to the method, model parameters and calculation complexity of multi-modal three-dimensional target detection are reduced, target detection efficiency and accuracy are improved, and the method has effectiveness and robustness in various complex scenes.

The embodiment of the invention provides an encryption method and device based on hidden code protection, electronic equipment and a storage medium, and the method comprises the steps: obtaining text information, an entropy source seed and a plurality of round keys, decomposing the text information, and determining a first text and a second text; performing encryption processing on the first text through the entropy source seed to obtain an intermediate ciphertext, and performing encryption processing on the second text through the entropy source seed and the round key to obtain a first entropy ciphertext; alternately performing shift operation and nonlinear transformation processing on the first entropy ciphertext to obtain a second entropy ciphertext; determining a new first text and a new second text according to the first text and the second entropy ciphertext, and encrypting the new first text and the new second text again; performing inverse transformation on the first text and the second text which are most encrypted to obtain a target ciphertext; according to the method, the nonlinear characteristic of the encryption algorithm can be enhanced on the premise that the algorithm performance is not increased, so that the encryption process is more complex, and the security of the ciphertext is improved.

According to the virtual power plant optimization method considering multivariate uncertainty under power-carbon market collaboration, a single virtual power plant carbon mechanism model is established, a plurality of virtual power plant operation optimization models are established, a plurality of virtual power plant power-carbon collaborative transaction models are established and solved, and finally the above steps are carried out for solving. According to the method, the influence of electricity price uncertainty and new energy output uncertainty of the electricity market on virtual power plant optimization scheduling can be effectively overcome, and the stable operation capability and the economic benefit of the system are improved; according to the invention, an electricity-carbon market collaborative decision-making mechanism is constructed, and economic low-carbon operation of the virtual power plant can be ensured while electricity-carbon balance is realized; according to the method, an electric energy sharing contribution degree and carbon quota sharing contribution degree model is constructed, profit distribution among virtual power plant alliance members is realized, and the fairness and enthusiasm of the members participating in cooperative transaction are improved; according to the invention, the ADMM algorithm is improved, the penalty factor is dynamically adjusted, the ADMM convergence speed is improved, and the influence of the initial selection of the penalty parameter on the algorithm performance is reduced.

The invention discloses a data leakage detection method and device, electronic equipment, a medium and a product. The method comprises the following steps: acquiring a to-be-detected file, and converting the to-be-detected file into a byte sequence; segmenting the byte sequence to obtain a byte block set; mapping the hash values of all byte blocks in the byte block set to corresponding bits of a cuckoo filter to obtain a feature fingerprint of the to-be-detected file; and calculating the similarity between the feature fingerprint of the to-be-detected file and the feature fingerprint of a preset reference data file, and judging whether the to-be-detected file has a data leakage risk or not according to the similarity. According to the technical scheme, various file formats can be efficiently identified, and on the basis of strictly controlling the false negative rate to guarantee the identification accuracy, the iterative improvement of the algorithm performance is further realized.

This invention relates to the field of information security technology and proposes a hybrid encrypted data transmission and storage method and system adapted to the domestically developed IT environment. The method includes: training an adaptive neural network based on the hardware information of the domestically developed CPU to output a capability feature vector; using a cache prediction network based on the capability feature vector to predict the performance of national cryptographic algorithms and output a performance prediction matrix; using a deep reinforcement learning network based on the performance prediction matrix to detect resource conflicts between combinations of national cryptographic algorithms, filtering the combinations, and outputting the optimal combination of national cryptographic algorithms; encrypting the data to be transmitted and outputting encrypted transmission data; transmitting the encrypted transmission data to the target node, decrypting the encrypted transmission data, re-encrypting and storing it, and outputting securely stored data. This invention improves the security and adaptability of data transmission and storage in the domestically developed IT environment, and meets the security compliance requirements of the domestically developed IT environment through a complete transmission verification and storage re-encryption process.

This invention discloses a design method and apparatus for a high-resolution direction-finding IP core for broadband signals based on HLS, relating to the field of signal processing technology. By constructing a comprehensive evaluation standard for sub-frequency wavelength and half-wavelength matching and sub-frequency energy—that is, a sub-frequency comprehensive evaluation function—this invention optimizes reference and candidate sub-frequency points that contribute significantly to algorithm performance. This effectively reduces the complexity of broadband signal focusing operations while ensuring algorithm performance. Furthermore, it integrates the high-resolution direction-finding algorithm for broadband coherent signals with a high-level synthesis implementation method, forming a high-resolution direction-finding IP core for broadband coherent signals that can be directly called by FPGAs, possessing strong versatility and reusability.

The invention provides a post-quantum federal learning system and a dynamic migration method, and relates to the technical field of network security. Dynamic switching between a basic cipher suite and a post quantum cipher suite is realized by introducing a dual-stack communication and cipher agility engine module, and continuous operation and business continuity of the system are ensured. The PQC performance optimizer overcomes the performance bottleneck of the PQC algorithm by reducing the calculation and communication overhead of the PQC algorithm in federated learning, so that the federated learning system can meet the real-time requirement. The hybrid encryption and hierarchical security aggregation module dynamically selects a security level according to a security policy, so that both security and efficiency are considered, and risks possibly existing in a single PQC algorithm are avoided. And the post-quantum security authentication and certification server verifies the integrity and authenticity of the hardware environment and the software environment of the client, a full-stack post-quantum trust chain is constructed, and the overall security of the system is improved.

The invention discloses a fitness terrain characteristic perception optimization algorithm selection method. A problem library and an algorithm library are constructed; the algorithm performance result of each problem is tested by using the algorithm, search trajectory data in the process is collected, and algorithm performance is ranked; based on the collected data, track features describing fitness terrain features are extracted, a plurality of features are combined into a high-dimensional track feature vector, and standardization processing is carried out to form feature-label pairing data; training a classification model and obtaining an explicit algorithm selection rule by taking the trajectory feature vector as an input feature and the algorithm performance ranking as a label; for a new optimization problem, inputting the feature vector into a trained algorithm selection rule model, and outputting a predicted optimal algorithm by the model; if the model does not definitely recommend a certain algorithm, randomly selecting an algorithm from an algorithm library as a solver; according to the scheme, the performance of the algorithm selection method is improved by intelligently sensing the characteristics of the fitness terrain, so that complex and changeable optimization requirements are more flexibly met.

This invention relates to a simulation verification method and system for a flexible direct-drive maglev line. The main components include constructing a virtual model of the flexible direct-drive maglev line; an FRT controller simulation module loading the virtual model and running a high-fidelity motor mathematical model and motion control algorithm; the FRT controller simulation module receiving control commands from an external process controller; real-time calculation based on the virtual model to generate dynamic response data of the virtual mover; and real-time feedback of the dynamic response data to the external process controller. This method completely eliminates the need for physical motors and the maglev line, allowing for comprehensive algorithm performance testing and process logic integration during the control program development phase, significantly shortening the development cycle.

The application discloses an education resource recommendation method fusing review text and review object content, first, a pre-training model XLNET of natural language processing is used to mine and encode the text, and shallow and deep feature vectors of users and education resources are obtained; second, interest migration of a user is learned and predicted by using LSTM and an attention mechanism, and output vectors of a user module and an education resource module are obtained; and finally, a multilayer perceptron is used to model nonlinear interaction behaviors between the user and the education resource, and a predicted score is obtained. Experimental results show that the fusion of the review text and the content data effectively solves problems such as cold start and data sparsity, improves the performance of the recommendation algorithm, more truly reflects the feature preferences of the user, and can further improve the recommendation accuracy.

The invention discloses a mine car dangerous driving behavior detection method and a detection system thereof. The detection method comprises the following steps: firstly, acquiring a real-time driving image of a mine car driver; and then target detection and state analysis are carried out on a detection target in the original image based on a YOLOv11 improved algorithm, and if a dangerous driving behavior is determined, a sound-light alarm is given out in the mine car, and meanwhile, a ground sound-light alarm is linked to realize double warning of sound and light. The detection system comprises a mining intrinsic safety type camera, a mining flame-proof and intrinsic safety type AI edge reasoning host, a mining intrinsic safety type wireless base station, a mine ring network, a ground dispatching host, a ground alarm and a mining intrinsic safety type audible and visual alarm, and has the advantages of full-dimension detection, excellent algorithm performance, dual early warning, adaptability to complex environments and the like. Dangerous driving behaviors can be accurately detected in real time, active early warning is achieved, and mine operation safety is improved.

The application relates to a distributed elastic fusion filtering method based on a generalized inverted pendulum system and relates to the technical field of information fusion filtering. A dynamic model of a generalized inverted pendulum system under a multi-sensor network with colored measurement noise is constructed, the dynamic model is converted into a dynamic model of a non-generalized inverted pendulum system, a local elastic filter is designed according to the measurement information of each node to filter the state of the system, a correction parameter matrix is calculated and substituted into the local elastic filter, the local state filtering of the generalized inverted pendulum system and the upper bound of the filtering covariance are obtained, and finally, an inverse covariance cross fusion method is used to obtain the best fusion filtering process. The application can effectively solve the fusion filtering problem of the generalized inverted pendulum system under the multi-sensor network affected by the colored measurement noise and the change of the filter parameters, can objectively fit the real engineering scene, is easy to realize online iterative calculation, and can improve the state estimation precision and the filtering algorithm performance.

The invention discloses a strapdown inertial navigation attitude algorithm precision evaluation method. A polynomial coefficient is set according to a required carrier dynamic environment; according to the polynomial, calculating attitude truth values of all time needing reference in an off-line manner; deriving an angular velocity analytical expression according to the attitude truth value, and generating gyroscope observation; the method comprises the following steps: taking gyroscope observation as input of a strapdown inertial navigation attitude algorithm to be evaluated, updating attitude parameters of a carrier, and obtaining an attitude calculation result; after each attitude algorithm is solved, the algorithm performance is quantitatively evaluated; after each algorithm performance evaluation, adjusting parameters and a coefficient matrix of the attitude trajectory model or storing current evaluation data; and storing the adjusted model parameters or evaluation data, and when a new high-order attitude algorithm needs to be verified or a maneuvering scene needs to be further expanded, turning to the initial step and sequentially carrying out recursive iteration until an evaluation task is completed. According to the method, large-angle maneuvering scene simulation and attitude / angular velocity analysis truth value acquisition can be considered, and the high-order attitude algorithm performance can be verified more accurately and efficiently.

The invention discloses a confidence-based multi-agent system interpretability enhancement method, system and device and a medium, and belongs to the technical field of artificial intelligence, and the method comprises the steps: collecting various information of each agent in a multi-agent system, and carrying out the preprocessing to obtain the preprocessed information; receiving the preprocessed information based on a deep learning model, and outputting a confidence value; correcting a confidence coefficient value in combination with environment change and information interaction between intelligent agents; a visual interface for displaying confidence is constructed, and the confidence calculation basis and time change of specific agent decision making are supported to be checked; the strategy of a multi-agent algorithm is adjusted according to the confidence coefficient of agent decision, the confidence coefficient is introduced in the training process to serve as a training target, and a cooperation mechanism between agents is optimized by utilizing confidence coefficient information. According to the method, the interpretability of the decision-making process of the multi-agent system is effectively improved, the algorithm performance is optimized, and the collaboration capability among the agents is enhanced.

Method for generating an optimized neural network architecture taking into account the hardware constraints of a hardware target. The present invention relates to a method for generating an optimized neural network architecture taking into account the hardware constraints of a hardware target (20), the method comprising: obtaining an algorithmic task (100); defining a neural search space (104) as a function of the algorithmic task (100); defining a hardware configuration space (102) of the hardware target; searching for a neural network architecture for the hardware target using a genetic algorithm (108) to optimize evaluation criteria (106), the evaluation criteria (106) being at least one hardware performance and at least one algorithmic performance;and evaluate the hardware performance(s) by using a performance predictor if the performance predictor has a reliability above a reliability threshold or the hardware performance(s) being determined by a performance evaluation by running the neural network on the hardware target (20) if the performance predictor has a reliability below the reliability threshold. Figure for the abstract: 2;

The present application belongs to the field of array signal processing, and relates to a switched antenna array DOA estimation method based on orthogonal matching pursuit joint tracking. In view of the problem of algorithm performance distribution imbalance in traditional switched antenna array DOA estimation, a switched antenna array DOA estimation method based on orthogonal matching pursuit joint tracking is provided. The present application takes a sparse reconstruction algorithm as a starting point, converts the DOA estimation problem into a sparse representation problem of a dynamic dictionary by constructing a spatially over-complete dictionary, utilizes the characteristic that the sparse solutions of various observation values correspond to the same support set, and jointly solves all sampling data by using an orthogonal matching pursuit joint tracking algorithm, so that the structural characteristics of the switched antenna array are fully utilized, the DOA estimation precision is improved, and high-precision and low-complexity switched antenna array DOA estimation is realized.

The application provides an atmospheric monitoring point distribution optimization identification method, device and equipment and a storage medium. It relates to the technical field of environmental monitoring. The method comprises the following steps: calculating Voronoi area and neighborhood density as node features based on the latitude and longitude of the site, combining with the AQI correlation to construct a graph structure; training a PollutionGNN model by using a double-layer GAT network with residual connection, establishing a three-objective optimization model of site number, coverage area and prediction error, and realizing the Pareto optimal solution set of monitoring efficiency by using an NSGA-III algorithm. The application solves the problem that the traditional point distribution method is difficult to balance the cost, coverage and precision, provides a theoretical basis for cross-regional application through the influence mechanism of geographical environment and pollution source structure on algorithm performance, finally realizes the upgrading of the monitoring network from experience layout to intelligent optimization, and provides an efficient and reusable solution for different cities.

The invention discloses a photovoltaic MPPT (Maximum Power Point Tracking) control algorithm, which relates to the technical field of photovoltaic control algorithms, utilizes the characteristics of simple structure, high convergence speed, high global optimization capability, wide applicability and the like of a gradient optimization algorithm, and improves the defects of premature convergence and sensitivity to control parameters of the gradient optimization algorithm. Meanwhile, the photovoltaic output energy efficiency and the stability of the photovoltaic output energy efficiency serve as the evaluation basis, an MPPT algorithm performance evaluation model is constructed through an analytic hierarchy process, photovoltaic MPPT performance is evaluated, and GTO performance is verified and improved.

The application discloses a kind of lightweight cryptographic algorithm security evaluation method and system, it includes: the algorithm information of lightweight cryptographic algorithm is acquired, and the strength evaluation value of static algorithm strength evaluation is obtained;Lightweight cryptographic algorithm simulation attack test is carried out, and the test evaluation value of its dynamic algorithm performance evaluation is obtained based on test result;Determine the energy performance coefficient of lightweight cryptographic algorithm test process, and the test correction evaluation value of its test evaluation value is corrected;Based on strength evaluation value and test correction evaluation value, the security evaluation value of lightweight cryptographic algorithm is determined, and its security level is determined based on security evaluation value.The application can accurately and comprehensively determine the security evaluation value of lightweight cryptographic algorithm by comprehensively considering the static algorithm strength evaluation value and the corrected dynamic test evaluation value, to determine its security level, provide scientific basis and guidance for its practical application, ensure the safety and reliability of system operation.