830 results about "Adaptive matching" patented technology

The invention discloses a digital twinborn model real-time linkage method driven by information physics fusion, particularly relates to the technical field of digital twinborn cooperative control of an industrial automation production line, and is used for solving the problems of instruction conflict and control failure caused by mismatching of virtual model parameters and dynamic capability of physical equipment in the prior art. According to the method, equipment state and material flow data are collected in real time, equipment dynamic degradation parameters are extracted to generate capability attenuation feature vectors, hidden process conflict path detection and resource preemption probability simulation are combined, a collaborative optimization model of equipment health and process scheduling is constructed, and a control instruction set containing dynamic capability constraints is generated. After the screening instruction is verified through physical constraint matching, parameters are corrected in a closed loop mode based on an execution result, a model is iterated, and self-adaptive matching of the virtual instruction and the physical equipment capacity is achieved; the reliability of the digital twinning control instruction and the stability of a production line are remarkably improved, and the risk of abnormal shutdown caused by an overrun instruction is avoided.

The invention relates to the field of image comparison, and discloses a multi-layer PCB alignment deviation detection system and method based on image comparison, and the method comprises the steps: obtaining high-resolution image data before and after lamination of a multi-layer PCB, carrying out the unified registration preprocessing of an original image through the combination of a multi-mode image fusion algorithm and a geometric distortion correction technology, and obtaining a multi-layer PCB alignment deviation detection result; constructing a standardized image registration input set; key alignment feature extraction is carried out on the image registration input set, and a multilayer structure graph model is constructed based on a graph neural network; in combination with the alignment reference map, dynamically adjusting an image comparison window and a search region by adopting a region attention mechanism and a local adaptive matching algorithm, and constructing an alignment deviation mapping map; constructing a deviation evolution model by using a time sequence behavior recognition network according to the constructed alignment deviation mapping graph and historical process deviation data; and performing comprehensive evaluation on the image registration input set based on the deviation evolution model and the generated early warning information. The method has the advantage of improving the accurate detection level.

The invention discloses a charge and discharge controllable system and method for a decommissioned battery, and relates to the technical field of intelligent charge control. By collecting the capacity fading rate, the internal resistance value, the cycle index and the environment temperature data of the decommissioned battery in real time, a health degree parameter is calculated by adopting a nonlinear coupling algorithm; and the future health degree evolution trend is predicted in combination with the LSTM neural network. And dynamically generating a grading label according to a preset scene threshold matrix, and matching the charging demand thermodynamic diagram with the battery grading label through a dynamic scheduling algorithm to realize intelligent distribution of charging and discharging power. And introducing a photovoltaic-battery-power grid cooperative power supply model, predicting and dynamically adjusting the power supply proportion based on the environment temperature and the photovoltaic output, and deploying to a target scene. And through a dynamic health degree evaluation and scene adaptive matching mechanism, the utilization rate of the retired battery is improved, the deployment cost of charging facilities is reduced, and the power supply reliability under multiple scenes is remarkably improved.

The invention discloses an intelligent scheduling method and system for a mobile robot, and relates to the technical field of robots. Extracting task basic attributes to form a static feature layer, establishing a dynamic feature layer formed by dynamic influence factors, and generating a double-layer task feature model; establishing a self-adaptive matching model of the robot and the task, performing feature fusion on the robot state data and the task feature matrix, and calculating a matching degree scoring matrix; a task allocation scheme is generated based on the matching degree scoring matrix, potential path conflict points are predicted, an avoidance strategy is determined, and conflict prediction and avoidance are executed; dynamically adjusting a task allocation and execution sequence according to the load state of the robot; and a performance closed-loop optimization framework is established, and scheduling strategy parameters are automatically adjusted. According to the method, the double-layer task feature model and the self-adaptive matching mechanism are constructed, and track prediction and conflict avoidance technologies are combined, so that intelligent matching and dynamic adjustment of the robot and the task are realized.

The invention relates to the technical field of sensors, and discloses a temperature compensation method based on a silicon resonance pressure sensor. According to the method, real-time temperature drift data is collected through a sliding time window, and a temperature sensitive area and a temperature non-sensitive area are accurately identified; establishing a coupling model of zero drift and full scale drift in the temperature sensitive area; then, piecewise linearization processing is conducted on the coupling model through a dynamic weight distribution strategy, and high-precision temperature compensation parameters are generated; and finally, according to the environment temperature value detected in real time, automatically calling the temperature compensation parameter of the corresponding segment, and carrying out real-time dynamic correction on the original output value of the sensor. According to the method, through region identification and model coupling of temperature drift characteristics, fine division and adaptive matching of a compensation strategy are realized, and the measurement precision and long-term stability of the silicon resonance pressure sensor in a wide temperature range are improved.

The invention provides a cross-modal adaptive matching online external parameter calibration method and system for a laser radar and a camera, and belongs to the technical field of online external parameter calibration of the laser radar and the camera, initial external parameters are obtained by using a static calibration result, a laser radar point cloud is projected to a visual angle of a virtual camera, an LIP image aligned with a camera image is generated, and a calibration result is obtained. Performing cross-modal segmentation on the LIP image and the RGB image, extracting a semantic mask, generating a cross-modal feature point pair through a strategy, and performing preliminary external parameter calculation in combination with a PnP algorithm; global optimization is carried out through a multi-modal loss function, and an inter-frame smoothing and increment optimization strategy is introduced to ensure accurate estimation of external parameters. Finally, the optimized external parameter matrix is applied to projection and matching of subsequent frames, and can adapt to environment change and data dynamic update in real time. The method has high precision and real-time performance, overcomes the defects that off-line calibration depends on a specific calibration target and cannot cope with dynamic environment changes, and adapts to application requirements in high-speed motion and complex environments.

The invention relates to a secure communication method for an edge node and terminal equipment based on dynamic key negotiation. The method comprises the following steps: the terminal equipment sends an initial signal when initiating a communication request; the edge node generates a scenarized first key negotiation parameter and feeds back the scenarized first key negotiation parameter after passing dual identity and state verification; after the terminal device decrypts the parameter, a dynamic entropy value is collected by combining a network adaptive sensor, and a second key negotiation parameter associated with the first parameter feature is generated; the two parties calculate session keys based on an ECDH algorithm, and the consistency of the keys is ensured through dynamic entropy verification and Hash comparison; and after the key negotiation succeeds, entering a hierarchical encrypted data transmission stage, and dynamically updating a session key based on a multi-dimensional trigger mechanism. According to the method, through hardware credibility verification, a scenarized key strategy, dynamic entropy enhancement and national secret algorithm adaptation, dynamic management of keys and adaptive matching of terminal resources are achieved, attacks such as equipment counterfeiting and parameter tampering are effectively resisted, and the high-security and compliance requirements in the fields of industrial control and the like are met.

The invention discloses an artificial intelligence self-adaptive education system based on big data, and relates to the field of intelligent education, the artificial intelligence self-adaptive education system comprises an education resource management module, a student data management module, a self-adaptive learning recommendation module, an auxiliary teaching decision module and a security reinforcement module, and a cloud data management library is used for classifying and integrating education resources. The method comprises the steps of establishing a knowledge graph, performing knowledge point classification on resources by constructing the knowledge graph, performing accurate matching on personal portraits of students and the knowledge graph through an incremental self-adaptive matching model, dynamically adjusting recommended learning resources in real time, analyzing collective learning problems through a big data intelligent analysis model, and generating collective teaching suggestions. A three-layer security firewall is adopted to ensure the security of the system; the problem that an existing self-adaptive education system neglects the learning progress and weak links of classes or groups is solved, and intelligent management, personalized learning recommendation and collective problem analysis of education resources can be achieved.

The invention relates to an intelligent interview scoring system capable of explaining decisions based on a large language model, and the system comprises a multi-mode resume analysis and feature coding unit, a resume feature adaptive matching unit, an interactive scoring and knowledge enhancement unit, and an answer quality evaluation unit. Text, image and audio features are extracted through a cross-modal attention mechanism of a multi-modal large language model, resume features are encoded into dynamic word vectors, and entity-level feature vectors are extracted; the post description text is encoded into a demand feature vector by a resume feature adaptive matching unit; calculating semantic similarity between the resume entity feature vector and the demand feature vector; the interactive scoring and knowledge enhancement unit dynamically retrieves knowledge fragments to generate a preliminary evaluation report containing a scoring basis; and the answer quality evaluation unit fuses the information density, the fluency and the integrating degree to generate a final score. And the whole-process intelligence from demand analysis to final decision making is realized.

The present invention relates to the technical field of power source power allocation. Disclosed are a direct current power source power allocation method and system for a generator status monitoring apparatus. The method comprises the following steps: by means of system data, constructing a variational problem model; solving the constructed variational problem model, and using a particle swarm optimization algorithm to optimize computing parameters of the variational problem model; and, on the basis of a computing result of the variational problem model, allocating the total output power of a hybrid energy storage system to an energy-type storage device and a power-type storage device according to a ratio. By means of using the convergence-guaranteed particle swarm optimization algorithm, the present invention not only excels in the solving speed but also shows significant advantages in computational accuracy; furthermore, by means of solving the variational problem model, more accurate allocation ratios are obtained; variational mode decomposition can achieve adaptive matching of the optimal center frequency and bandwidth for each mode, thus effectively separating intrinsic mode components and achieving frequency domain partitioning of signals.

The invention relates to the technical field of large model algorithms, in particular to a knowledge distillation-based field large model lightweight adaptive method and system, and the method comprises the steps: obtaining knowledge distillation parameters and student model performance parameters, and building a nonlinear mapping relation between the knowledge distillation parameters and the student model performance parameters; the optimal parameter combination is optimized and solved based on the mapping relation, and target knowledge distillation parameters are generated; issuing the target parameters to a training engine, monitoring performance deviation in real time and triggering re-optimization; in the reasoning process, performance fluctuation is monitored, and model characteristics are managed and controlled; target domain data characteristics are collected, a mapping relation is corrected in combination with big data analysis, and the domain adaptation capacity is improved; a knowledge base and a case base of historical distillation data are constructed, a standardized adjustment scheme is formed, and self-adaptive matching is achieved. According to the scheme, through precise modeling, dynamic optimization, real-time monitoring and knowledge reuse, the knowledge distillation efficiency, model robustness and field adaptability are remarkably improved, and systematic technical support is provided for large model lightweight.

The invention provides a palm vein feature extraction and depth adaptive matching method, which relates to the technical field of biological recognition, and is characterized in that a three-dimensional space-time convolutional neural network and a graph neural network are combined to extract space depth features, time blood flow dynamic features and blood vessel topological features of palm veins, and information bottleneck is utilized to optimize and screen a feature subset with the most distinguishing property; the method breaks through the limitation of traditional two-dimensional texture feature extraction, obtains abundant palm vein information through multi-source data acquisition (such as near-infrared imaging, a light field camera and a spectrum sensor), overcomes the interference of environmental factors such as illumination, angle and noise, achieves the high-precision modeling and stable extraction effects of palm vein features, and can achieve the real-time extraction of the palm vein features in a complex scene. A high recognition rate is kept under different illumination conditions or user hand posture changes, the robustness of the palm vein recognition system is remarkably improved, and the method is suitable for scenes with high safety requirements.

The invention provides a meteorological-power collaborative prediction method and device, and belongs to the field of new energy power systems. The method provided by the invention comprises the steps of determining condition parameters of a prediction task and a prediction site, wherein the condition parameters at least comprise multi-source meteorological observation data and new energy station operation data; according to the multi-source meteorological observation data and a pre-trained meteorological prediction large model, generating a future time sequence meteorological prediction result of the prediction site in a future time period; according to the time scale of the prediction task and historical weather conditions, adaptively matching a corresponding parameter set of the pre-trained power prediction large model; and inputting the future time sequence weather prediction result into the large power prediction model to generate a new energy power prediction result of the new energy station. According to the meteorological-power collaborative prediction method and device provided by the invention, the technical problems that the collaboration of meteorological prediction and power prediction in a new energy power system is insufficient, the precision is limited, and a complex scene is difficult to adapt can be solved.

The invention discloses an intelligent scheduling method for video equipment based on solar power supply, and relates to the technical field of energy management and video. Through multi-dimensional data acquisition, intelligent prediction and dynamic scheduling technologies, the energy management challenge faced by solar power supply video equipment in remote areas is effectively solved, the power generation amount, equipment power consumption and environmental change data are acquired in real time by using an Internet of Things sensor, and the energy management efficiency is improved by combining time sequence analysis and a machine learning algorithm. The power generation amount and power consumption fluctuation are accurately predicted, an energy supply and demand matching model is established based on a dynamic planning algorithm, a power distribution strategy is optimized, operation of high-priority equipment is guaranteed preferentially when the power generation amount is insufficient, meanwhile, the working mode of the equipment is dynamically adjusted through a reinforcement learning algorithm, and energy consumption optimization and self-adaptive matching are achieved. The energy utilization efficiency and the system stability are improved, the adaptive capacity to dynamic environment changes is enhanced, and continuous and stable operation of video equipment under complex conditions is ensured.

The invention discloses an interactive multi-model underwater maneuvering target tracking method and system based on a pure azimuth second-order EKF. The method comprises the following steps: establishing a discrete state space model of a pure azimuth target tracking system; performing second-order linearization on the pure azimuth measurement model; calculating mixed input of a sub-filter corresponding to each model in the model set of the interactive multi-model algorithm; estimating a target state through parallel filtering of the azimuth-pure second-order EKF sub-filter corresponding to each model in the model set; updating the probability of each model in the model set; and combining the estimation results of all the sub-filters to obtain a final estimation result of the target state. According to the method, the target tracking precision can be improved, the problem of highly nonlinear measurement of the UUV on a target detection and tracking system through a passive sonar is effectively solved, good real-time performance is achieved, and real-time tracking of the UUV on the motion state of the underwater maneuvering target is achieved through adaptive matching of the real motion of the target through the interactive multi-model algorithm.

The invention relates to the field of computer vision, in particular to an image feature recognition method based on computer vision, which comprises the following steps of: preprocessing an input image, removing image noise and correcting image gray deviation to obtain a preprocessed image; performing multi-scale feature extraction on the preprocessed image, obtaining image local texture features and global contour features under different scales, performing dynamic weight fusion, calculating dynamic fusion weight based on a feature response value and a local complexity factor, and performing dimension reduction processing on fusion features through a principal component analysis algorithm to obtain a target feature set; and matching the reference features, calculating the Euclidean distance between the target feature set and the reference features, determining an adaptive matching threshold, judging a matching result, and completing image feature recognition. According to the method, through adaptive filtering and gray level equalization preprocessing, Gaussian pyramid multi-scale feature extraction and dynamic fusion and adaptive threshold matching, feature extraction accuracy, characterization capability and recognition robustness are improved.

The invention discloses a horizontal five-axis turning and milling composite center with a self-adaptive matching system. The horizontal five-axis turning and milling composite center comprises a tool turret, a boring tool mechanism, a tool changing area and a control system. A radio frequency identification reading device is arranged in a tool changing area of the tool turret and the boring tool mechanism, and an RFID chip tag is attached to the outer surface of each tool and used for storing tool serial numbers, specifications and length compensation parameters. And the control system realizes self-adaptive matching and sequential error correction of the cutter and the working procedure through radio frequency identification. The tool changing area is provided with a visual monitoring assembly which comprises a plurality of sets of industrial cameras, an adjustable light source and a structured light stripe illumination module and is used for detecting the clamping posture, the abrasion state and the type of a tool. The system has self-learning and digital twinning simulation functions, and can simulate the motion trail of the cutter and predict the interference risk before the cutter is changed. According to the method, integrated control over automatic tool recognition, posture correction, abrasion detection and path prediction is achieved, and the tool changing precision, the operation stability and the automation level of the horizontal five-axis turning and milling composite center are remarkably improved.

The invention belongs to the field of query optimization, and particularly relates to a fragmentation storage and query optimization method and system for a high-concurrency database, and the method comprises the steps: analyzing business features through a preset decision tree model, and selecting an optimal fragmentation key, constructing a fragmentation rule engine and initializing a database in combination with hash, range and list fragmentation rules and a fragmentation splitting-merging strategy; generating a query abstract syntax tree by using an ANTLR4 analyzer, and detecting whether a preset Cube is hit or not to directly return a result; if not, dynamically routing to a target fragment list based on a fragment key field, load balancing and a failover strategy, accelerating data retrieval in combination with a high-frequency index, and combining fragment-level results through a T-TopK algorithm; according to the method, the self-adaptive matching of the fragmentation strategy and the service requirement, the calculation push-down of the query process and the result optimization are realized, and the low delay and the high availability in a high-concurrency scene are ensured.

The invention discloses a self-adaptive sound effect management method and system with a scene analysis module, and relates to the field of self-adaptive sound effect management, and the method comprises the following steps: S1, collecting environmental acoustic characteristics and user behavior data in real time through the scene analysis module; s2, dynamically extracting user hearing preference parameters and scene classification labels; s3, constructing a virtual audition sweet zone matched with the physical space; and S4, dynamically adjusting the phase delay and the energy distribution proportion of each frequency band, and generating an optimal loudness equalization curve meeting the multi-scene constraint. According to the method, adaptive matching of sound field characteristics is realized in a complex dynamic scene through a multi-level perception-optimization closed-loop architecture, acoustic feature extraction and physical space modeling are deeply fused, the technical boundary of traditional pure signal processing is broken through, and the problem of multi-target conflict is solved based on cross-modal collaborative learning and a global optimization algorithm.

The invention relates to the technical field of image precision optimization, in particular to a surveying and mapping image mathematical precision evaluation method based on dense point cloud matching, which comprises the following steps: acquiring and preprocessing image data, synthesizing an image gradient modulus length and a maximum principal curvature, and introducing a surface normal vector partial derivative to obtain a surface normal vector partial derivative; calculating a curvature-driven depth discretization step length, and generating an initial point cloud data set; constructing a point cloud complex network structure diagram and calculating the topology durability of a point cloud topology structure to obtain an optimized point cloud data set; based on the optimized point cloud data set, in combination with the target image information, the adaptive curvature gradient weight and the error diffusion control value, constructing an adaptive matching cost function, and calculating a point cloud matching cost matrix; and calculating an optimal point cloud transformation matrix based on the point cloud matching cost matrix. According to the surveying and mapping image mathematical precision evaluation method based on dense point cloud matching, low-texture region point cloud matching error suppression and structure optimization of high-curvature region point cloud on a complex curved surface are realized.

The invention provides a multi-modal data feature alignment and optimization system and method based on an FPGA, and the system comprises an input buffer module which is used for receiving multi-modal data, carrying out the data processing, and optimizing the buffer depth configuration based on the modal number and the data rate; the space-time perception scheduler is used for extracting space-time characteristics of the multi-modal data and fusing the space-time characteristics according to dynamic weights; the self-adaptive matching kernel is used for mapping the multi-modal feature codes to a unified semantic space and ensuring semantic consistency; the reconfigurable feature compression unit is used for dynamically adjusting a compression ratio through tensor CP decomposition and controlling a reconstruction error based on a rank parameter; a self-calibration optimization mechanism is adopted, feature alignment errors are monitored in real time, mapping matrix parameters are dynamically adjusted through a gradient descent algorithm, and calibration frequencies are switched according to data distribution changes. Through the FPGA pipeline architecture and self-adaptive resource allocation, the calculation delay is reduced, and the throughput is improved.

The invention belongs to the field of power distribution automation, and relates to a power grid equipment frequency adaptive resampling method based on a dynamic scale factor, and the method comprises the steps: carrying out the point compensation through employing a linear interpolation method when a power grid equipment detects that the number of continuous lost points is smaller than or equal to 3; selecting a voltage signal channel or a current signal channel to carry out zero crossing point detection and real-time frequency calculation; when the real-time frequency obtained through calculation deviates from the nominal frequency, the power grid equipment introduces a dynamic scale factor for resampling, sampling points are increased when the real-time frequency is higher than the nominal frequency, and sampling points are reduced when the real-time frequency is lower than the nominal frequency; writing data obtained by resampling into a calculation annular cache region of the power grid equipment; and performing FFT calculation on the data in the calculation annular cache region to obtain power grid parameters. The sampling interval of the data acquisition unit is dynamically adjusted by increasing or reducing the sampling points, so that the sampling time sequence and the actual signal period are synchronized, the sampling frequency and the real-time frequency are adaptively matched, and the sampling distortion problem under the power grid frequency fluctuation is solved.

The invention relates to the technical field of image processing, in particular to a communication image filtering and denoising method, which comprises the following steps of: calculating a phase consistency numerical value of each pixel point in a communication image based on the input communication image, selecting an area as a structural point according to the phase consistency numerical value, and extracting a local direction attribute of the structural point, and generating a local structure feature map. According to the method, the anisotropic wavelet basis parameters are dynamically generated based on the local structure feature map, so that the wavelet function is adaptively matched with the image region characteristics in shape, direction and scale, the high-frequency detail retention capability is improved, and the problem of frequency band aliasing of a fixed wavelet basis in a non-uniform noise scene is solved. A dynamic adjustment threshold matrix is constructed in combination with local structure features and wavelet coefficient distribution, noise and effective signals are distinguished through signal saliency evaluation, wavelet coefficient amplitude is adjusted in a targeted manner, and excessive smoothness is avoided while noise is suppressed.

The invention relates to the technical field of unmanned aerial vehicle path planning methods, in particular to an unmanned aerial vehicle path planning method based on an improved leapfrog algorithm. According to the method, dynamic weight adjustment information is obtained by carrying out dynamic adaptation matching analysis on flight stage information and environment threat information, a flight stage weight mapping matrix is established, and fitness function weights are dynamically adjusted according to core targets in take-off, cruise and landing stages; weight correction is carried out in combination with environmental data such as obstacle grid density and a meteorological threat coefficient, so that the algorithm focuses on core constraints in the current stage, and target conflicts caused by global unified weight are avoided; according to the method, environmental threats are sensed in real time by means of multi-sensor fusion, when abnormal states such as over-high obstacle density and strong wind are detected, an algorithm is triggered to be suddenly stopped and restarted, a population is reconstructed according to the current position, threat avoidance weight is enhanced, the defect that an existing mechanism can be updated only after the current iteration period is completed is overcome, and the method is suitable for large-scale popularization and application. And the response time to an abnormal environment is shortened.

The invention relates to the technical field of three-dimensional reconstruction, and provides a local feature adaptive fusion and high-fidelity new view angle synthesis method based on three-dimensional reconstruction. According to the method, a local feature module is introduced into a point cloud network, robust matching of local features is realized, a self-adaptive matching strategy is combined, high-resolution images can be efficiently processed, and the precision and stability of three-dimensional reconstruction are remarkably improved. The method is suitable for application scenes such as three-dimensional rendering, three-dimensional reconstruction and new view angle synthesis. The specific process comprises dense reconstruction and initial rendering; and carrying out missing region repairing and iterative optimization. Generating a high-precision point cloud based on an improved stereo matching algorithm, and rendering an initial image of a target view angle in combination with camera parameters; a conditional video generation model is adopted, and holes, distortion and artifacts in initial rendering are repaired; and carrying out progressive complementation on the occlusion region by using the view angle consistency constraint, and improving the reconstruction integrity by jointly optimizing the point cloud geometry and texture. According to the method, the final picture quality is remarkably improved through feature matching optimization and efficient processing capacity. Local feature branches and global context information are coordinated, and the problem of matching ambiguity in a complex scene is solved; the adaptive fusion strategy gives consideration to high-resolution image processing efficiency and detail retention; and the detail fidelity and the overall quality of the rendering result are effectively improved.

The present invention relates to a targeted sensing-oriented twin substructure interaction method and system and the use. The method comprises: acquiring multi-element inspection and monitoring data and finite element influence line data of a main structure; on the basis of the inspection and monitoring data and the finite element influence line data of a non-key area, using a sparsity adaptive matching pursuit algorithm to solve a boundary condition of a key area, so as to achieve data fusion of a first level; for the key area, establishing a refined twin substructure finite element model taking into account the effect of structural degradation and, on the basis of the inspection and monitoring data and the finite element influence line data of the key area, correcting the material attribute of the refined twin substructure finite element model to achieve data fusion of a second level; and calculating a correction force on the basis of the boundary condition of the key area and the material attribute, and using the correction force as an equivalent external load to act on a node of a global finite element model, thus achieving the interaction between a substructure and the main structure. The present invention achieves mutual fusion and common communication between the monitoring data and the finite element model and the interaction between each substructure and the main structure.

The invention relates to the technical field of voice synthesis, in particular to a digital human tone adaptive matching method and system based on voiceprint feature migration. According to the method, firstly, a deep voiceprint decoupling network is adopted, a user voice sample is processed through a multi-task adversarial learning strategy, pure timbre identity embedding decoupled from content and general acoustic features is extracted, and meanwhile, multi-scale timbre features used for representing timbre macroscopic features and microcosmic details are extracted. Then, according to user configuration, a combination is selected from the multi-scale timbre features, voice synthesis weights are distributed, and a customized voiceprint feature set is formed; and finally, the pure timbre identity is embedded and the customized voiceprint feature set is used as a condition to be injected into a diffusion flow matching nerve vocoder, and the vocoder combines probability flow conversion and gradual denoising refining to generate a digital human speech synthesis waveform which has unique timbre of the user and corresponds to the content according to the to-be-synthesized content information. The invention aims to realize accurate, flexible and highly personalized adaptive matching of digital human timbre.

The invention discloses an ultra-harmonic rapid detection method based on compressed sampling and multiple measurement vectors, and solves the problems of high sampling rate, low reconstruction efficiency and insufficient spectrum resolution in the prior art. Sub-Nyquist compressed sampling is realized through a sparse random matrix, and the hardware burden is reduced; a dynamic step size sparse self-adaptive matching pursuit algorithm (DSSAMP) is adopted to reconstruct signals, the step size is adjusted in a self-adaptive mode in combination with residual energy changes, and the convergence speed and precision are balanced; an improved multi-measurement vector model (IMMV) is constructed through framing processing and introduction of interpolation factors and prior frequency point screening, and the calculation efficiency is improved; and finally, extracting super-harmonic frequency, amplitude and phase parameters based on an orthogonal matching pursuit (OMP) algorithm. According to the method, the high resolution is ensured, the sampling rate and the calculation complexity are remarkably reduced, and the real-time requirement is met.

The invention relates to the technical field of tunnel and underground engineering, in particular to a method for dynamically calculating the jacking resistance of a jacking pipe based on mud real-time sensing data driving. Comprising the following steps: S1, constructing a pipe-slurry-soil interface friction force database; s2, calculating the real-time detected jacking resistance of the mud; s3, optimizing a friction resistance database based on data driving; and S4, the oil cylinder thrust is calculated in real time based on the jacking resistance. According to the method, the pipe-slurry-soil contact state is directly judged by detecting the thickness and distribution of resistance reduction slurry in real time, and real-time, accurate and unitized calculation of the jacking resistance is realized by calling a dynamically optimized interface friction database; a machine learning technology is utilized, so that database parameters can be self-learned and self-adaptively corrected according to field measured data, and the long-term accuracy of a calculation model is ensured; based on resistance distribution calculated and predicted in real time, an intelligent control system is constructed, and self-adaptive matching and regulation of thrust of the pushing oil cylinder and the hinged oil cylinder are achieved.

An AI auxiliary image annotation method based on semi-supervised learning comprises the following steps: acquiring a to-be-annotated image, extracting multilayer features of the image by using a deep learning network, and generating a feature map; the global context information and the local feature information are fused, and enhanced feature representation is generated and used for contour prediction; on the basis of enhanced feature representation, modeling a relationship between vertexes by adopting a graph convolutional network, and constructing a dynamic graph structure; according to the dynamic graph structure, calculating self-adaptive matching weights among the vertexes to realize a dynamic vertex pairing strategy; predicting an initial contour position by combining global and local features through a dynamic vertex pairing strategy; based on the predicted initial contour, a deformable convolutional network is adopted to realize progressive optimization of the contour; in the progressive optimization process, vertex position information is updated in real time, and the graph structure relation is adjusted; and finally, outputting an accurate object contour labeling result, and completing an image labeling task.