19 results about "Maximum a posteriori estimator" patented technology

The application discloses a MRI paravertebral muscle segmentation method based on Bayesian fusion and probability shape modeling, and comprises the following steps: a) a Fourier-Gaussian process probability shape model is established to parameterize the muscle shape, the Fourier-Gaussian process probability shape model comprises radial modeling based on a Fourier basis function to represent a muscle cross-section profile, and axial modeling based on a Gaussian process to capture the axial variation law of the muscle shape; b) a muscle edge feature model based on a convolutional neural network is constructed to perform multi-scale feature extraction on the muscle edge in the MRI image; and c) a Bayesian segmentation framework is designed, and based on a maximum posterior estimation principle, online segmentation of the paravertebral muscle in the MRI image is realized. The application is customized according to the characteristics of the paravertebral muscle segmentation task through the idea of fusing high-quality artificial labeling and a powerful learning model, so that the application can achieve high segmentation and reconstruction quality under a small amount of interaction, and has strong shape change description capability.

The invention discloses a method for quickly generating high-resolution carbon emission raster data, which comprises the following steps of: fusing national carbon emission panel data and night light remote sensing images, establishing a provincial-scale logarithmic regression model, extracting parameters of the logarithmic regression model and constructing a parameter characteristic matrix; a space-time constraint K-Means algorithm is adopted to perform clustering analysis on provincial administrative units throughout the country and divide the provincial administrative units into K carbon emission type areas, for each carbon emission area, an area regression model is established by using logarithmic carbon emission and night light data, and a clustering center parameter is used as Gaussian prior to obtain a clustering result; and performing parameter updating through Bayesian maximum posteriori estimation, estimating the carbon emission by using the model, performing correction on a pixel scale by adopting a zero error method, and finally generating a national carbon emission grid data set. The space-time constraint clustering and Bayesian maximum posteriori estimation method is introduced on the basis of carbon emission modeling, and balance of few models and high precision is achieved.

The invention discloses a quantum image sensing method and device based on exposure modulation. The sensing end performs frame-by-frame exposure modulation on the same scene in continuous K frames, and regulates and controls the expected number of incident photons of each frame; the pixels output a 1-bit binary image in each frame and obey Poisson distribution. And the collection end takes K frames as a group to carry out maximum posteriori estimation to obtain reference intensity estimation of each pixel. In order to improve robustness and speed, an RED-PRO iterative algorithm based on FISTA acceleration is adopted in the reconstruction process, after gradient descent of a data consistency item, the data is projected to a pre-training de-noising device manifold, and a pixel adaptive step length is given based on Fisher information. According to the method, the signal-to-noise ratio is remarkably improved and artifacts are inhibited in a scene with low illumination, saturation and limited dynamic range, the 1-bit sensing circuit architecture does not need to be changed, and the existing QIS / SPAD sensor and the lens module can be compatible only through exposure modulation and software reconstruction.

The application discloses a kind of structural damage identification bayesian model fast solution method based on zero sample transfer learning, belong to structural damage identification and machine learning technical field.To solve the problem of generalization performance decline of structural damage identification bayesian model.The application includes constructing the bayesian model of structural damage identification;Combined with sparse auto-encoder and adversarial learning, construct and train the adversarial sparse auto-encoder, the adversarial sparse auto-encoder includes source domain sparse auto-encoder module, target domain sparse auto-encoder module and domain discriminator module;PhyCNN proxy model is constructed;Source domain label data set is generated, and source domain PhyCNN proxy model is constructed;Design parameter freezing transfer strategy, construct target domain PhyCNN proxy model;Based on target domain PhyCNN proxy model, sample posterior sample, update structural damage identification bayesian model using posterior sample, calculate the maximum posterior estimation value of structural damage parameter, realize the positioning and quantization of structural damage.

The invention belongs to the technical field of pseudo satellite positioning, discloses an indoor pseudo satellite signal carrier-to-noise ratio positioning method based on a commercial smart phone, and solves the core problem through innovative design. The invention provides a positioning technology based on Bayesian reasoning alternating optimization dynamic estimation. According to an attenuation relationship between a carrier-to-noise ratio observation value and a distance base station, maximum posteriori estimation (MAP) is calculated through Bayesian reasoning, and an alternating estimation strategy is adopted to iteratively optimize an attenuation coefficient and a position until the attenuation coefficient and the position converge, so that the change of the attenuation coefficient is effectively dealt with, and the performance difference between smart phone devices is relieved; compared with a traditional weighted centroid method, the average positioning precision is improved by 30%. The technology provided by the invention is not only suitable for smart phones with different performances, but also can be used. The technology provides a universal solution for high-reliability positioning of the smart phone under the same-frequency interference condition.

The invention belongs to the technical field of spaceflight and space situation awareness, and particularly relates to a space debris short arc data association method and device based on a factor graph, and the method comprises the steps: obtaining optical angle measurement short arc data from an observation station, and carrying out the preprocessing; an angle observation initial orbit determination algorithm is adopted to obtain an orbit state initial value; constructing a cost matrix, performing preliminary screening on track state initial values corresponding to different short arc observation data, and generating a limited number of multi-correlation hypotheses through disturbance sampling expansion; constructing a nonlinear factor graph under each association hypothesis; solving the maximum posteriori estimation, and selecting an optimal association hypothesis according to a residual error and an information criterion; and outputting the associated cluster, the track state and the uncertainty thereof, and carrying out consistency check. According to the method, efficient and robust multi-arc-segment association and preliminary orbit estimation can be performed on optical angle measurement short arc data in foundation and space-based observation scenes, and scene self-adaption, real-time performance and high accuracy are realized.

This invention discloses an online flow field estimation path planning method based on Bayesian inference, aiming to solve the problem of efficient detection and estimation of unknown flow fields by mobile robots. The method first discretizes the two-dimensional flow field region into a grid and constructs a probabilistic roadmap. In each planning round, a weighted maximum a posteriori (MAP) estimation problem is constructed based on the residuals of velocity measurements and fluid equations to solve for the flow field state and calculate the posterior accuracy matrix. Candidate paths are generated through bundle search and then prospectively expanded using subsequent information gain. After mapping path nodes to a set of proposed measurements, a joint prediction covariance matrix is ​​calculated. Based on a task weight matrix dynamically weighted according to velocity magnitude, a task-weighted posterior uncertainty reduction score is calculated for the candidate paths. The path with the highest score is selected for execution, and new measurements are collected. This process is repeated until a preset termination condition is met. This method integrates physical constraints and prospective search, guiding priority exploration of high-velocity regions and significantly improving the efficiency of online robot perception.

The application provides a digital twin method and system for realizing bone growth, which is suitable for the technical field of digital twin of orthopedics. The method comprises the following steps: acquiring two-dimensional tomographic image data of a fracture site, establishing a three-dimensional surface geometric model of the fracture site, and forming a finite element model of the fracture site; based on the finite element model, assembling a solid mechanics, pore seepage and biological reaction coupled solver, and performing multi-physical field coupled solving; periodically acquiring follow-up image data of bone fracture healing, realizing online data assimilation of modulus, and outputting a rehabilitation load scheme; the application uses variational assimilation based on maximum a posteriori estimation to fuse the observation modulus field generated by the follow-up image and the simulation result, realizes continuous convergence of the model and the real state, and on this basis, real-time batch calculation of different load parameter combinations is realized, and the rehabilitation load scheme meeting the safety constraint and the optimal healing speed is output.

This invention discloses a method for bridge expansion joint status identification and early warning based on Bayesian data fusion, comprising: S1, constructing a disaster-causing factor system and a failure cause database for expansion joints; S2, collecting multi-dimensional data of expansion joints; S3, preprocessing and extracting features from the collected multi-dimensional data; S4, constructing a Bayesian network model based on a fault tree; S5, fusing multi-dimensional data to determine the prior probability of the Bayesian network model; S6, constructing a conditional probability table for each node of the Bayesian network based on maximum a posteriori estimation and the analytic hierarchy process; S7, quantifying the failure probability of expansion joints through forward inference using the Bayesian network, classifying risk levels and risk warnings; and screening the main disaster-causing factors through backward inference using the Bayesian network. This invention integrates multi-dimensional data and multi-disciplinary theoretical methods, solving the problems of single data dimensions, low fusion efficiency, and insufficient early warning accuracy in existing technologies, and significantly improving the accuracy of expansion joint damage identification and the reliability of risk assessment.

This invention discloses a high-precision positioning and mapping method, system, and storage medium based on FMCW velocity, comprising: S2, acquiring point cloud data and inertial data; S3, estimating the FMCW velocity of the corresponding frame's moving platform; and S4, performing pre-integration fusion to obtain the first... k +1 frame relative to the first k FMCW-IMU pre-integration model for frame motion; S4, obtain the first frame. k Frame and the k +1 frames of residual term; S5, minimize the residual term between frames. k Frame and the k S6. Sum of the squares of each term in the residual term between +1 frames to obtain the maximum a posteriori estimate of the corresponding frame; k The maximum a posteriori estimate of the frame is obtained. k +1 frame of pose parameters to form the first k Frame to the k +1 frame of transformation matrix, and the newly acquired first frame k The point cloud data of frame +1 is transformed using a transformation matrix and then added to the map to complete the reconstruction. This method can solve the problem of accumulated errors caused by inertial navigation drift, and improve the positioning accuracy and robustness of the map.

This invention relates to the field of bridge digital twin technology, specifically to a dynamic update method for bridge digital twins based on multi-fidelity learning and sequential Bayesian inference. It constructs a parameterized finite element model and trains a multi-fidelity deep neural network surrogate model using high-fidelity and low-fidelity simulation data, replacing expensive simulations at extremely low cost. During online updates, modal features are extracted from monitoring data, and the posterior of the previous time step is used as the prior. The probability distribution of parameters is sequentially updated using Bayes' theorem, and finally, the digital twin is updated with the maximum a posteriori estimate. This invention overcomes computational bottlenecks through a multi-fidelity surrogate model, achieves probabilistic quantification and dynamic reduction of parameter uncertainty through sequential Bayesian inference, and verifies the physical consistency of the surrogate model and constructs an information-based prior through interpretable AI technology, thereby constructing an efficient, probabilistically robust, and physically reliable dynamically evolving digital twin.

The invention provides a multi-sensor fusion train positioning method and system based on factor graph optimization, and belongs to the technical field of train positioning. 16-dimensional state variables are constructed to serve as factor graph optimization input variables; constructing an INS pre-integral factor; when the attitude change of the train exceeds a set threshold value, selecting the laser radar frame at the moment as a key frame, discarding the other laser radar frames in the two key frames, and constructing a local map by adopting a sliding window; constructing a LiDAR odometer factor; constructing a TLC factor; and adding the INS pre-integration factor, the LiDAR odometer factor and the TLC factor into the factor graph, calling iSAM2 for optimization based on maximum posteriori estimation, and outputting a train pose estimation result under the ENU coordinate system. According to the method, the factor graph model is constructed, and the trackside features such as the transponder, the kilometer post and the hectometer post are utilized, so that high-precision positioning of the train in the satellite navigation signal limited and denial environment is realized.

The application discloses a noise image smoothing denoising method and relates to the technical field of computer digital image processing. The method comprises the following steps: in a Bayesian probability framework, a Gaussian probability model is used to model noise, and a gradient l0 norm and an l1 norm are used to model a clear image; three regularization constraint coefficients are introduced, the three models in the step 1) are weighted and summed, and a noise image restoration problem model is constructed; a quadratic penalty function method is used for solving, two auxiliary variables corresponding to horizontal and vertical gradients of the clear image are introduced, a penalty coefficient is introduced, and the noise image restoration problem model is converted. In a Bayesian maximum posterior estimation framework, the clear image is modeled by introducing the l0 norm and the l1 norm of the image gradient, the regularization constraint condition is formed, and an optimal estimation problem is comprehensively constructed, and an iterative optimization algorithm is designed, so that the noise can be effectively suppressed, and the significant edges can be sharpened and enhanced, thereby generating a high-quality restored image.

The invention relates to the field of intelligent operation and maintenance of a remote sensing data processing system, and discloses an intelligent abnormal root cause positioning method and system for system operation and maintenance, and the method comprises the steps: obtaining a system topology, determining a component dependency relationship, and building a dynamic fault tree model based on the system topology to describe a fault logic dependency relationship of a hardware layer, a network layer and an application layer; the fault tree is converted into a Bayesian network, node prior probability distribution is trained by using a historical fault library, and hardware indexes, network parameters and application logs collected in real time are fused to construct an evidence chain; and calculating a Bayesian posterior probability by adopting a maximum posterior estimation (MAP) algorithm, and dynamically outputting a root cause probability sequence. According to the method, the problem that the monitoring system is large in automatic alarm information amount and complex in system structure, so that abnormity checking is difficult is solved, the accuracy and timeliness of abnormal root cause positioning are improved, rapid recovery of system faults can be supported, and thus the system stability is improved.

The application discloses a quantum image sensing method and device based on exposure modulation. The sensing end performs frame exposure modulation on the same scene in K continuous frames, and controls the expected number of incident photons of each frame; the pixel outputs a 1-bit binary image in each frame and obeys a Poisson distribution. The acquisition end performs maximum a posteriori estimation on K frames as a group to obtain a reference intensity estimation of each pixel. In order to improve robustness and speed, the reconstruction process adopts a RED-PRO iterative algorithm based on FISTA acceleration: after the gradient descent of the data consistency term, the projection is performed to the pre-trained denoiser manifold, and the pixel adaptive step is given based on the Fisher information. The method significantly improves the signal-to-noise ratio and suppresses artifacts in low-light, saturation and dynamic range limited scenes, and does not need to change the 1-bit sensing circuit architecture, and only needs to be compatible with the existing QIS / SPAD sensor and lens module through exposure modulation and software reconstruction.

One or more embodiments of the invention provide a reservoir lithofacies identification method, apparatus and device, a storage medium and a program product. The method comprises the following steps: constructing a likelihood probability model of lithofacies according to well seismic data of a verification well, wherein the likelihood probability model is used for predicting the likelihood probability of the lithofacies; according to the occurrence frequency of each lithofacies in the verification well, a transition probability model of the lithofacies is constructed by using a Markov chain model, and the transition probability model is used for predicting the transition probability of the lithofacies; combining the likelihood probability model and the transition probability model, and utilizing Bayesian maximum posteriori estimation to construct an objective function for identifying lithofacies; constructing a recursive score function for the target function, and dynamically solving the score function through a recursive process to obtain an optimal lithofacies of each position; and after recursion is completed, path backtracking is carried out according to the optimal lithofacies at the final position, and an optimal lithofacies sequence is obtained to serve as a lithofacies recognition result.

The application relates to a footprint positioning method and system for a full waveform spaceborne laser altimetry system. In the method, firstly, a transmission waveform, reference terrain data, an original footprint sequence and a received waveform are collected; a simulation waveform is generated based on the transmission waveform and the reference terrain data, and a waveform similarity heat map is created by using a Pearson correlation coefficient based on the received waveform and the simulation waveform; secondly, a profile matching error map is generated by using a root mean square error based on the original footprint sequence and the reference terrain data; then, matching probabilities of the waveform similarity heat map and the profile matching error map are calculated respectively, a posterior probability is calculated by using a Bayesian theory based on the two matching probabilities, a fusion probability map is generated, and a maximum posterior estimation in the fusion probability map is taken as an optimal offset; finally, the original footprint sequence is corrected by using the optimal offset, so that the footprint positioning is completed. Compared with the prior art, the application has the advantages of improved positioning accuracy and robustness.

The application provides a kind of tide correction coastline datum determination method, belong to coastline detection technical field, the present application is extracted by collecting high-resolution satellite image and airborne laser radar data input based on attention guide's neural architecture growth algorithm constructed sea-land boundary identification model refinement instantaneous coastline, establishes the nearshore tidal numerical model containing shallow water partial tide term and composite partial tide term, obtains nearshore refinement tidal forecast tidal level sequence by variational data assimilation optimal control algorithm, establishes spatial distribution function by correlating refinement instantaneous coastline with corresponding forecast tidal level, and the coastline position corresponding to the lowest tide level is obtained by bayesian maximum posterior estimation method by comprehensively processing multi-temporal remote sensing data as the coastline datum, solve the technical problem that the accuracy of coastline datum determination is influenced by tidal dynamic change and water density difference, resulting in inaccurate spatial positioning.