11 results about "Maximum a posteriori estimation" patented technology

The invention discloses a high-precision positioning mapping method and system based on FMCW speed, and a storage medium. The method comprises the following steps: collecting point cloud data and inertial data; s2, estimating the FMCW speed of the mobile platform of the corresponding frame; s3, performing pre-integration fusion to obtain an FMCW-IMU pre-integration model of the (k + 1) th frame moving relative to the kth frame; s4, obtaining a residual term between the kth frame and the (k + 1) th frame; s5, minimizing the sum of squares of all items in a residual term between the kth frame and the (k + 1) th frame, and obtaining the maximum posteriori estimation of the corresponding frame; and S6, according to the maximum posteriori estimation of the kth frame, obtaining a pose parameter of the (k + 1) th frame to form a conversion matrix from the kth frame to the (k + 1) th frame, converting the newly obtained point cloud data of the (k + 1) th frame through the conversion matrix, and adding the converted point cloud data into the map to complete reconstruction. According to the method, the problem of accumulative errors caused by inertial navigation drift can be solved, and the positioning precision and robustness of mapping are improved.

A method of receiving operation data about a collection of machines. The operation data characterizes one or more aspects of the operation of at least one machine of the collection of machines. The method further includes establishing a first conjugate prior or a second conjugate prior for a first distribution probability density function. The method further includes performing, based on the operation data and the first and second conjugate priors for the first distribution probability density function, a Maximum A Posteriori (MAP) estimation to determine distribution parameters for the first distribution probability density function. The data are samples of the key performance indicator. The method further includes predicting, based on the distribution parameters, a probability the key performance indicator will take a value for a number of the machines. In addition, the method includes scheduling maintenance for the machines based on the probability.

The application relates to the technical field of intelligent vehicle management, and discloses a vehicle state monitoring method and system based on deep learning, wherein the vehicle state monitoring method based on deep learning comprises the following steps: based on structural mechanics, constraint analysis, vibration anomaly detection, loose positioning and fatigue evaluation monitoring task physical dependency relationship, a task dependency directed graph is generated; a multi-modal monitoring data matrix is input into a shared encoder of a multi-task learning network to generate a unified representation vector; the unified representation vector is respectively input into a vibration anomaly detection decoder, a loose positioning decoder and a fatigue evaluation decoder to output initial prediction results of each task; a conditional probability model between task outputs is established based on a constraint relationship defined by the task dependency directed graph, and a consistent diagnosis result meeting physical constraints is generated through maximum posterior estimation. The application solves the technical problems of one-sided diagnosis results, mutual contradictions and low calculation efficiency.

The application provides a cross-view splicing detection method based on coarse-to-fine matching, which comprises the following steps: performing feature extraction on a to-be-detected image to obtain feature points; setting a coarse matching threshold, matching the feature points to obtain a coarse matching set, wherein the coarse matching set comprises correct matching pairs and incorrect matching pairs; modeling the coarse matching set to obtain a mixed model, performing maximum a posteriori estimation on the mixed model, distinguishing the correct matching pairs from the incorrect matching pairs, and obtaining an output correct matching pair set; and positioning a splicing tampering region of the to-be-detected image according to the correct matching pair set to obtain a detection result.

The application discloses an infrared small target adaptive mask generation method and product based on physical priori, and the method comprises the following steps: acquiring an infrared image to be processed and a single-point label located at any position in a small target effective response area; taking the single-point label as a seed point, performing polarity unification processing on the infrared image according to local background statistical information of the seed point neighborhood; establishing a candidate small target area with the seed point, constructing a priority queue of pixels to be expanded, and initializing the candidate area and a statistical quantity; modeling a target mask generation problem as a maximum posterior estimation problem, and constructing a posterior energy function; gradually expanding the candidate area based on the priority queue and a greedy search strategy, and recording the posterior energy in the expansion process to obtain an energy record sequence; selecting an energy peak value from the energy record sequence, and generating a target pixel-level mask according to a corresponding optimal state backtracking; and extracting geometric supervision information according to the target pixel-level mask.

This invention relates to an LTE / IMU integrated navigation method using an adaptive covariance factor graph. This method models the navigation problem as a global maximum a posteriori estimation problem by tightly coupling and integrating the factor graph, achieving optimal estimation of navigation pose information. The main factor nodes include IMU factors and pseudorange factors. The cost function consists of the difference between the observed and estimated quantities, enabling time-varying analysis and correction of state error and pseudorange error values. Simultaneously, a probability transfer model for sensor-acquired information is constructed, and a detailed optimal estimation solution is provided. Furthermore, addressing the problem of the covariance matrix not changing over time in traditional factor graph integrated navigation methods, this invention proposes an adaptive covariance method, which can effectively improve positioning accuracy.

The present invention discloses a line feature matching method based on Fourier representation and posterior probability estimation, including: detecting and extracting line features on an image sequence pair, calculating binary descriptors; obtaining initial line feature matching pairs; introducing a compact Fourier series to represent the mapping relationship satisfied between adjacent images; establishing a prior maximum likelihood estimation; iteratively obtaining the maximum a posteriori estimation parameters satisfied by the line feature matching pairs; establishing a uniform distribution probability model; screening the image line feature matching pairs that satisfy the probability threshold; searching for the remaining feature matching pairs, and iteratively completing and realizing the accurate and robust matching of the image line features. The present invention solves the problem of accurate and robust matching of image line features only through basic line feature descriptors, represents the mapping relationship between images through Fourier, combines the geometric relationship satisfied by the line feature matching pairs to establish a likelihood estimation model, and realizes the accurate and efficient matching of line features based on the iterative convergence of the EM algorithm.

The invention provides a high-latitude transfer alignment algorithm based on a factor graph, and belongs to the technical field of high-precision navigation. The invention aims to solve the problem that the traditional Kalman filtering method is difficult to deal with the strong nonlinearity and parameter coupling problem of high-dimensional regional navigation, so that the precision and robustness of transfer alignment are reduced. The method comprises the following steps: S1, modeling a high-latitude transfer alignment problem into a time sequence factor graph model; s2, calculating the posterior distribution probability of the system according to the time sequence factor graph model; s3, according to the posterior distribution probability, obtaining an error state estimation value of the sub inertial navigation relative to the main inertial navigation through maximum posterior estimation; and S4, correcting the navigation parameters of the sub inertial navigation according to the error state estimated value, and completing transfer alignment in the high-latitude environment.

This invention belongs to the field of aerospace and space situational awareness technology, specifically relating to a method and apparatus for associating short-arc data of space debris based on factor graphs. The method includes: acquiring and preprocessing optical angle measurement short-arc data from observation stations; obtaining initial orbital state values ​​using an angle observation initial orbit determination algorithm; constructing a cost matrix to initially screen the initial orbital state values ​​corresponding to different short-arc observation data, and generating a limited number of multiple association hypotheses through perturbation sampling expansion; constructing a nonlinear factor graph under each association hypothesis; solving for the maximum a posteriori estimate, and selecting the optimal association hypothesis based on residual and information criteria; outputting the association cluster, orbital state, and its uncertainty, and performing a consistency check. This invention enables efficient and robust multi-arc association and preliminary orbit estimation of optical angle measurement short-arc data in both ground-based and space-based observation scenarios, achieving scene adaptability, real-time performance, and high accuracy.

The present invention provides an improved dynamic harmonic estimation method and system, comprising: collecting harmonic voltages of synchronous phasor measurement devices of each branch or bus of a power grid; substituting the harmonic voltages into a nonlinear state equation and a measurement equation model; using an improved Sage‑Husa square root unscented Kalman filter algorithm, and estimating the harmonic current of the branch to be estimated based on the nonlinear state equation and the measurement equation model, to obtain an estimated value of the harmonic current of the branch to be estimated; wherein the improved Sage‑Husa square root unscented Kalman filter algorithm includes: updating the state of the nonlinear state equation and the measurement equation model based on measurement update and forgetting factor. The present invention addresses the problem that accurate estimation results cannot be obtained from unknown noise when using square root unscented Kalman filter to estimate the harmonic state of a power system. The present invention introduces the idea of ​​maximum a posteriori estimation based on the Sage‑Husa filter algorithm and utilizes the forgetting factor of the Sage‑Husa filter to achieve real-time dynamic estimation of unknown noise and obtain high-precision harmonic estimation results.