39 results about "Data space" patented technology

The application relates to an industrial process state monitoring method based on robust space-time joint projection, which comprises the following steps: data preprocessing and standardization; data space manifold structure and time dynamic information capturing: calculating the similarity degree between samples according to neighborhood overlapping order similarity, obtaining a data similarity matrix, and thus obtaining the overall space manifold information of process data; subsequently, calculating a data time derivative matrix, extracting slow change characteristics in the data, and obtaining key dynamic information; feature extraction and monitoring model building based on robust space-time joint projection: comprehensively integrating space-time information, performing dimension reduction projection on the data, obtaining key low-dimensional characteristics for retaining the representation of system operation states, constructing monitoring statistics and monitoring thresholds based on the key low-dimensional characteristics, and building an offline process monitoring model; real-time online monitoring of the state of an industrial system.

The application provides a splitting implementation method of a RISC-V vector cryptography instruction, which splits a message scheduling instruction into a plurality of micro-instructions for execution on a processor with a vector data path bit width of 128 bits according to a preset splitting rule in an instruction decoding stage, the splitting rule comprising: obtaining a message scheduling instruction in a SHA-512 algorithm, the message scheduling instruction being used to expand an input message block into a required message sequence; splitting the message scheduling instruction into a 0th type micro-instruction for processing a transformation path of sig0 and a 1st type micro-instruction for processing a transformation path of sig1 according to a calculation type; and the method does not introduce additional storage data space, and can reduce the area overhead of the processor for implementing the instruction while maintaining the unit data path operation throughput.

The application discloses a millimeter wave radar point cloud compression method based on heterogeneous representation and implicit neural network. First, the point cloud fine geometric features are extracted through a point-based compression network, and then the global semantics are aggregated through a voxel-based compression network to realize point-voxel heterogeneous joint representation and preliminary compression. Secondly, an adaptive octree is constructed to encode the compressed data space structure to generate a partition bit stream, and a context-aware entropy coding is used to statistically compress the voxel-level high-dimensional features to generate a feature bit stream. Then, the partition bit stream, the feature bit stream and the neural network weight for implicit representation are jointly transmitted. Finally, the coarse-grained point cloud is decoded and recovered at the receiving end, and is input into a point-based synthesis network based on implicit neural representation together with the neural network weight, and through continuous surface modeling and detail compensation, a high-fidelity point cloud is finally reconstructed. The method significantly improves the compression efficiency and reconstruction quality of the millimeter wave radar point cloud at a low code rate.

The application provides a perception method based on a light-weight geometric information decoupling graph convolution point cloud perception model, and belongs to the technical fields of point cloud perception and deep learning. The method comprises the following steps: obtaining candidate region retrieval by discretizing the original point cloud data space coordinates through 2D voxel down-sampling based on the light-weight geometric information decoupling graph convolution point cloud perception model; obtaining adjacent edge index by using a scale factor-based expansion KNN algorithm based on the candidate region. The encoder performs multiple graph convolution and down-sampling processes based on the adjacent edge index and the point cloud subset, and the decoder performs multiple up-sampling and graph convolution processes based on the point cloud subset and the adjacent edge index. The application can complete point cloud classification by extracting multi-layer graph convolution features through the encoder. For point cloud segmentation and target identification tasks, the decoder gradually reconstructs spatial details in the up-sampling stage, effectively compensating for the loss of geometric information caused by down-sampling in the encoding process.

The application provides an optical remote sensing physical adversarial sample generation method based on a multi-task loss function optimization, and the method comprises the following steps: step one, data and model preparation; step two, arranging a parameterized physical adversarial sample in a training set image; step three, training the parameterized physical adversarial sample; step four, arranging the optimized parameterized physical adversarial sample in a test set image, and testing the adversarial performance by using a test model; and output: a physical adversarial sample design drawing; a series of physical adversarial samples are prepared according to the proposed generation method, and experiments are conducted in a data space and a physical real world to test the interference ability of the physical adversarial samples; the experimental results fully prove the effectiveness of the proposed generation method, and the generation method has wide application value.

The application discloses a kind of flexible production line-oriented multi-robot vision cooperative positioning and grabbing method, it is related to flexible production line robot vision cooperative technical field, comprising: deployment includes industrial camera, 3D laser radar distributed vision node and wrist camera, cover operation area and guarantee field of view overlap and real-time feedback;Multi-source vision data space-time calibration, control synchronization error, unified coordinate and optimize registration accuracy, introduce temperature compensation correction parameter;Cooperative detection tracking workpiece, share results and dynamically adjust tracking mode;Plan global optimal grabbing pose, generate candidate pose and screen;Cooperative control robot motion trajectory, predict avoidance interference;Through vision servo correction pose deviation, trigger grabbing and confirmation, complete transfer.The application improves positioning and grabbing precision and efficiency, enhances multi-robot cooperativity, reduces interference, adapts production line workpiece replacement demand, guarantees long-term operation stability, provides strong technical support for flexible production line.

This invention provides a method and system for monitoring energy consumption anomalies in steel production processes based on data space. The method includes: calculating the spatial correlation strength between variables and quantifying the influence weight score of each variable on energy consumption indicators; constructing a three-dimensional multi-layer data space integrating time, space, and energy consumption indicator dimensions; extracting time-series data and inputting it into SAISAE; embedding the influence weight score into attention weight calculation to obtain energy consumption weighted time features and reconstructing them to generate a first reconstructed variable; inputting each variable individually into MISAE; extracting the spatial features of each variable; combining this with a mutual information adjacency matrix constructed from the spatial correlation strength to obtain spatial correlation features characterizing the collaborative correlation properties of each variable; reconstructing these spatial correlation features to generate a second reconstructed variable; constructing monitoring statistics separately; and performing decision-level fusion of the two types of monitoring statistics using a Bayesian fusion algorithm to output the final monitoring result of energy consumption anomalies in the steel production process. This invention can perform energy consumption anomaly monitoring.

The application provides a kind of water environment multi-index data space-time field establishment and early warning method, comprising the following steps: S1, detection data acquisition: obtain environmental detection data;S2, time segmentation: based on detection time information, environmental detection data is divided in multiple time slices;S3, uniform grid construction: based on the coordinate position of all detection points, construct uniform two-dimensional space grid;S4, spatial interpolation reconstruction: environmental index rasterization space-time field is generated on two-dimensional space grid using spatial interpolation method;S4, standardization processing: based on the average and standard deviation of interpolation result calculation standardization processing, to dimensionless rasterization space-time field;S6, multivariate two-dimensional function construction: based on the dimensionless rasterization space-time field of each environmental index, construct multivariate two-dimensional function data object;S7, multivariate function principal component analysis, obtain spatial characteristic function, principal component score and variance contribution;S8, baseline modeling and anomaly detection, output abnormal early warning information.

The application relates to a traffic flow data space-time completion method and system in a big data environment, which comprises the following steps: collecting road vector data, point of interest and surface of interest information and real-time multi-source sensing data, and performing data preprocessing; converting the road vector data into a weighted directed graph; establishing a spatial correlation relationship between the point of interest and surface of interest information and road sections; matching and fusing the weighted directed graph and the multi-source sensing data; extracting features by using an attention mechanism to form a preliminary feature vector; further analyzing to obtain a spatial feature vector and a time feature vector representing the global spatial correlation of the road network, and inputting the spatial feature vector and the time feature vector into a fusion prediction model to obtain a preliminary traffic flow prediction value; and based on the preliminary traffic flow prediction value, dynamically calculating the traffic distribution proportion of each intersection, and iteratively calculating the movement and accumulation of vehicles among road sections to output a final traffic flow prediction value. Compared with the prior art, the application has the advantages of adapting to special scenes and being strong in model robustness.

This invention provides a method, device, equipment, medium, and product for predicting dam deformation at multiple monitoring points and providing early warning of safety zones based on a spatially weighted graph structure. It relates to the field of artificial intelligence technology. The method includes: acquiring deformation data and environmental data from multiple monitoring points of the target dam; constructing a spatial correlation weight matrix and graph structure for the target dam based on the spatial coordinate information of each monitoring point; and inputting the deformation data, environmental data, spatial correlation weight matrix, and graph structure into a pre-trained dam deformation prediction model to obtain the dam deformation prediction result at future times. Through this method, the model can accurately simulate the spatiotemporal coordinated changes in the dam deformation process, thereby generating more accurate dam deformation prediction results, improving the model's prediction accuracy and robustness, and enabling high-precision prediction of dam deformation, which is beneficial for subsequent dam risk early warning and operation and maintenance optimization.

This application provides a method and related apparatus for optimizing training data of a multi-vertical category model. The method includes: constructing a full vocabulary and a word attribute library for each vertical category; performing multi-dimensional value evaluation on each word in the full vocabulary based on the word attribute library to obtain the comprehensive value of each word; optimizing the full vocabulary based on the comprehensive value of each word to obtain a high-value vocabulary for the vertical category; and merging and deduplicating the high-value vocabulary for each vertical category to obtain a hybrid vocabulary. In this way, the comprehensive value of words is quantified through their attribute information, thereby optimizing the independent vocabulary for each vertical category, filtering out low-value words, retaining high-value words, and finally merging and deduplicating the high-value vocabulary for each vertical category to obtain a hybrid vocabulary. This reduces the vocabulary size required for model training, significantly compresses the training data space, and lowers training costs.

The application discloses a kind of slope intelligent monitoring method and system, method includes: through the original data of slope surface layer and deep layer physical mechanics parameter acquisition by multiple type sensor array;Effective data is obtained by eliminating outliers through edge computing node;Effective data space-time correlation analysis and supplement missing data form complete data set;Input slope stability evaluation model calculates stability coefficient;According to the safety level of coefficient determination, generate risk warning signal when lower than threshold value;The results are transmitted to remote monitoring center and stored to form historical archives.The system includes multiple type sensor array deployment unit, edge computing data processing unit, space-time correlation data analysis unit, slope stability evaluation unit, safety level determination and early warning unit, remote monitoring and data storage unit, each unit cooperates to realize slope global real-time monitoring, accurate evaluation and graded early warning, provide reliable technical support for slope safety prevention and control, applicable to various engineering slope monitoring scene.

The application discloses a compound game reinforcement learning method and device for a quadrotor unmanned aerial vehicle cluster, and relates to the technical field of unmanned aerial vehicle cluster control. The method comprises the following steps: directly approaching value evaluation and cooperative strategy in the "state-control input" data space of cluster flight by using the "evaluation-execution" network of the unmanned aerial vehicle, avoiding repeatedly solving the high-dimensional coupled HJ equation set under the strong coupling nonlinear dynamics of the unmanned aerial vehicle, and thus significantly reducing the modeling accuracy requirement and the calculation cost. By introducing a filtering mechanism with a "forgetting factor", the pose / velocity / relative formation error, control instruction, energy consumption and safety cost and other samples collected in the historical flight process of the cluster are subjected to weighted integral processing, a compound "evaluation" network error and a regression signal fusing current and historical information are constructed, and a limited incentive criterion that can be online tested is given at the information matrix level, so that the unmanned aerial vehicle cluster can still maintain effective learning under the condition that the continuous PE condition is not met.

The present application relates to a kind of electromagnetic reversing valve life prediction method based on flow signal, belong to hydraulic component life prediction field.The present application utilizes improved lumped average modal empirical decomposition method, by adding positive and negative pairs of noise to reduce the degree of modal aliasing in modal decomposition, using permutation entropy to detect abnormal component, realize the accurate adaptive modal decomposition of nonlinear measured signal;Application kernel principal component method, introduce nonlinear function as kernel function, based on the principle of mapping, convert original space into high-dimensional space, form new data set, using principal component analysis for data dimension reduction in high-dimensional data space, form feature vector, get performance degradation fusion index;Through cubic exponential smoothing processing;Finally based on the trained adaptive neural network model, establish the life prediction model of electromagnetic reversing valve, calculate the life of electromagnetic reversing valve.This method can effectively predict the pressure drop trend and life of electromagnetic reversing valve.

In some implementations, a device may receive, from a data stream, a set of data points arranged in a dimensional data space. The device may compare the set of data points to identify one or more clusters using values of a distance parameter for data points included in the set of data points, wherein the values of distance parameter includes different values of the distance parameter for different data points. The device may transmit an indication of the one or more clusters to cause a device to display information associated with the one or more clusters. The device may receive, from the device, feedback information associated with at least one data point, wherein the feedback information indicates that at least one data point is associated with an error. The device may modify a value of the distance parameter associated with the at least one data point to a modified value.

This invention belongs to the field of smart agriculture and precision cultivation management, and relates to a method and system for dynamic management of crop growth cycles based on big data. The method includes: collecting initial environmental data and expected target yield data of the target field to construct an extreme value tensor representing resource constraints throughout the entire growth period; acquiring feedback monitoring data after the preceding growth period and extracting state feedback parameters reflecting the cumulative resource input; determining the remaining resource weights for subsequent growth periods based on the state feedback parameters, and using these remaining resource weights to slice the extreme value tensor to generate a data state space representing the dynamic feasible region; and calling an optimization prediction operator to restrict the parameter search path within the data state space, outputting predicted trajectory data and determining dynamic management instructions. This invention converts resource input into data space constraint factors, causing the predicted trajectory to converge towards the physical safety boundary, mapping the physical compression of later paths by earlier execution deviations.

This invention relates to the field of medical image 3D reconstruction and fusion technology, specifically a method and system for multimodal medical image fusion and 3D reconstruction, comprising: receiving and standardizing raw image data streams from different devices; extracting intensity and texture features of each modality of image, constructing a feature space mapping relationship, and generating a fused volume data field based on this using a feature-level fusion strategy; after filtering and enhancing the volume data, extracting isosurface geometric meshes and simplifying and smoothing them to obtain a preliminary 3D model surface; mapping the preliminary surface back to the volume data space, and performing surface detail optimization and topology correction; and finally outputting a 3D reconstructed model integrating complementary information from multiple modalities. This invention improves the information integration and geometric accuracy of the reconstructed model by performing deep fusion in the feature space and employing a closed-loop optimization mechanism of surface-volume data back-mapping.

The application discloses an adaptive course learning training method for an electromagnetic target identification model and belongs to the technical field of artificial intelligence. The method aims at the problems of poor generalization ability and slow convergence speed of an existing deep learning model in a complex electromagnetic environment, and a hierarchical training data space based on signal-to-noise ratio and environmental interference characteristics is constructed. By initializing a deep neural network model and using dynamic feedback factors containing a task difficulty coefficient, model accuracy and a false detection penalty term, an adaptive weighted loss function is constructed. In the training process, the weight parameters are adjusted in real time according to the model convergence state to control the migration of the model from the basic feature domain to the complex adversarial domain. Through the quantitative stage migration control and weight adaptive redistribution mechanism, the application effectively solves the problems of gradient disappearance and catastrophic forgetting in multi-target detection, significantly improves the identification accuracy and robustness of the model in a low signal-to-noise ratio and strong interference environment, and is suitable for military reconnaissance and unmanned aerial vehicle defense systems.

The application discloses a kind of atmospheric pollutant estimation system and method based on geographic weighted depth forest, it is related to environmental monitoring, atmospheric science and machine learning cross technical field.The application breaks through the shallow application of traditional geographic weighted random forest to geographic weight, first complete multi-source heterogeneous data space-time matching and preprocessing;Geographic weight loss random forest (GLRF) is constructed, generates geographic weight by adaptive kernel function and reconstructs split criterion by embedding loss function in depth, realizes model innovation in combination with adaptive bandwidth and continuous space-time feature coding.Using Monte Carlo resampling to weaken meteorological interference, generate high spatio-temporal resolution meteorological normalized PM2.5 data nationwide;The application solves the problem of insufficient modeling accuracy of non-stationary space at national scale, overcomes the defects of poor adaptation, low fitting and difficult to identify mutation of traditional model, can eliminate meteorological fluctuations, capture human pollution mutation such as heating, and provide support for environmental protection policy evaluation and clean energy transformation.

The application provides an adaptive power inspection robot and an inspection method, and belongs to the field of intelligent robots. The robot comprises: a multi-modal perception data space-time normalization fusion module for generating a multi-dimensional device perception tensor; a device and task characteristic quantization module for calculating a device health quantization value, an environmental operation difficulty coefficient and a task urgency quantization value; a robot body capability state coding module for generating a robot body capability state vector; and a coupling field gradient decision module for inputting the above-mentioned quantization values, the state vector and features dynamically extracted from the perception tensor into a pre-trained coupling field model, outputting a global behavior gradient field, and decoupling into a four-wheel chassis motion instruction and a mechanical arm operation instruction. The application realizes deep real-time cooperation of movement and operation in the inspection process, solves the problems of low efficiency, slow response and poor adaptability of traditional inspection robots caused by the fragmentation of movement and operation, and improves the accuracy and safety of intelligent inspection of substations.

Embodiments of the present disclosure provide a power transformer multi-modal image data space-time synchronization method and device. The method comprises: collecting multi-modal image data of a power transformer; determining one kind of modal image data from the multi-modal image data as first modal image data, and determining other modal image data as second modal image data; performing time synchronization on the first modal image data and the second modal image data according to respective time stamps of the first modal image data and the second modal image data; calculating a coordinate transformation matrix from the second modal image data to the first modal image data according to the first modal image data and the second modal image data at the same time after time synchronization, and performing space synchronization on the first modal image data and the second modal image data after time synchronization based on the coordinate transformation matrix. In this way, the space-time synchronization effect of multi-modal image data can be improved.

The application provides a track traffic multi-system information interaction control method based on train-ground network cooperation, relates to the track traffic control technical field, and comprises the following steps: obtaining train-mounted equipment and ground infrastructure operation data, extracting a semantic feature vector and realizing heterogeneous data standardization, executing train-ground data space-time alignment to establish a synchronous state set, constructing a directed graph structure to identify a key state subset, further constructing a multi-objective optimization function to solve a Pareto optimal solution set, and deriving an optimal cooperative control instruction sequence. The application realizes information interaction and cooperative control among multiple systems, and improves track traffic operation safety and energy utilization efficiency.

The application discloses a kind of multi-source heterogeneous data space target identification method based on mamba module, to solve the precision and generalization low problems caused by the single modality data limitation, multi-task coupling risk and data sample deficiency in existing space object identification.The core steps of the method include:1, data preprocessing and fusion input;2, mamba module feature extraction and prediction: target ID classification and subject / sailboard / load quantity prediction can be independently executed respectively;3, data enhancement processing;4, robust training optimization, high confidence clean data is filtered by DivideMix Gaussian mixture model, and the model is iteratively optimized by combining the greedy pseudo-label generation strategy;5, target identification and out-of-library discrimination, based on the maximum probability principle, output target category and component quantity, and combine entropy enhancement to improve the discrimination accuracy of out-of-library targets.The application effectively avoids the traditional multi-task all or nothing risk, and improves the recognition accuracy and generalization performance of the model in complex scenarios.

A method for rapidly detecting a straight line, a plane and a hyperplane in a multi-dimensional space. The new method has two important advantages: firstly, the model corresponds to a total least square fitting algorithm and has better tolerance to data noise, so as to solve the problem of the precision of detecting a target on a parameter space segmentation line by means of fast Hough transform being too low; and secondly, in the integrated fast Hough transform, targets that are close to each other in a data space are gathered together in a parameter space, and a calculation process in the parameter space can be displayed by using a visual graph, thereby rapidly determining the number of targets, guiding the setting of system parameters, and distinguishing a target that is repeatedly recognized.

This invention discloses a Köppen climate identification and classification method based on geomorphic features. It acquires multi-source geomorphic feature data and utilizes bilinear resampling and Gaussian smoothing techniques to achieve spatial normalization and terrain denoising preprocessing. Seventeen feature parameters are extracted to construct a geomorphic feature dataset. A terrain roughness index is constructed using the local standard deviation of elevation to divide complex and flat terrain regions. Spatial iterative clustering with variable-density seed points is performed to generate geographically homogeneous superpixel units. A training set is constructed through visual interpretation of sample points. A stacked ensemble learning framework is used to generate base layer predicted features and train a meta-classifier layer to complete the fusion result feature set. Global superpixel classification is performed for backfilling, and the results are topologically merged and extracted based on spatial adjacency principles to generate maps. This method achieves detailed identification and classification of Köppen climate at the hundred-meter level, providing a robust, convenient, and economically feasible solution for accurate local climate identification in complex geographical environments.