12results about How to "Address underutilization" patented technology

The invention discloses a vehicle body grid parameter dynamic optimization method and system based on model simulation, and belongs to the crossing field of computer-aided engineering and artificial intelligence technology. According to an existing vehicle body grid optimization scheme, deep mining and analysis of grid data cannot be achieved, and consequently the accuracy and reliability of a grid optimization result are not high. According to the vehicle body grid parameter dynamic optimization method based on model simulation, a vehicle body historical database module, a vehicle body grid optimization module and a domain knowledge base module are utilized to match vehicle body grid data to be optimized with vehicle body historical data to obtain the most similar vehicle body historical data; the information is used as experience information and verification basis of vehicle body grid parameter optimization; and then experience information and a verification basis are reused, the to-be-optimized grid data of the vehicle body are analyzed and processed, and optimization suggestion information is generated, so that historical data can be deeply mined and analyzed, and the accuracy and reliability of a grid optimization result can be effectively improved.

This invention proposes a method and system for monitoring electricity trading risks based on a multi-source signal mechanism and a sparse hybrid expert model. The method includes: collecting historical day-ahead electricity market price data and raw data on influencing factors; performing data preprocessing to construct a multi-source fusion signal from price spread signals and bidding space signals; using a random forest algorithm for feature selection; constructing a set of similar day data based on the bidding space signal and incorporating it into the training set; using a sparse hybrid expert model as the model framework, with the price spread signal and bidding space signal as targets, training different prediction models; and searching for an optimal declaration curve based on the Northern Eagle optimization search algorithm by predicting the price spread signal and bidding space signal, using the short-term power prediction curve as a benchmark. This invention effectively solves the problem of insufficient model generalization ability, realizes risk monitoring in electricity spot trading, and improves the accuracy of electricity spot trading risk assessment while enhancing the stability of the power system.

The invention discloses an oil gas recovery intelligent control system and method based on multi-source data fusion, and relates to the technical field of intelligent control. The system comprises a data acquisition and classification module, an influence intensity judgment module, a priority dynamic adjustment module, a multi-source fusion control module and an equipment execution module. The data acquisition and classification module divides direct and indirect data, the influence strength judgment module dynamically captures data association strength without a preset rule, the priority dynamic adjustment module matches data priorities according to strength sorting, the multi-source fusion control module fuses high-priority data to generate a control instruction, and the equipment execution module adjusts operation parameters. Through dynamic collection of classified data, real-time judgment of influence intensity, dynamic adjustment of priority, multi-source fusion control and adaptation to multi-medium complex working conditions, the control accuracy is improved, the high recovery rate is stably maintained, emission reaches the standard, and the operation and maintenance risk is reduced.

The application discloses a kind of nonlinear carrier phase removal methods and systems of fringe projection image.The method obtains the fringe projection image of first and second view angle with nonlinear carrier;Including the following steps: (1) the fringe projection image of first and second view angle is fused using image splicing algorithm;(2) using singular value decomposition, the first principal component obtained is regarded as background component, and other components are regarded as foreground component;(3) detect foreground region and background region;(4) polynomial curved surface fitting is carried out to the background region of first and / or second view angle, and the first and / or second nonlinear phase is obtained;(5) subtract respective carrier phase and the respective nonlinear phase obtained in step (4).The application obtains global fusion fringe projection image without obstruction by double camera fringe imaging system, and detects background region based on singular value decomposition, so that nonlinear carrier phase is accurately removed.

The invention provides a large-scale generation method and system for interpretable question and answer pairs for the vertical field, and relates to the technical field of data processing, and the method comprises the steps: 1, constructing a cross-modal heterogeneous meteorological knowledge graph; heterogeneous graph neural network modeling is carried out on the cross-modal heterogeneous meteorological knowledge graph, an embedded vector of each node is learned, a knowledge graph with node embedded vectors is obtained, semantic community division is carried out on an entity based on the embedded vectors, and an optimized knowledge graph is generated. According to the method, through knowledge graph optimization, conical semantic space construction, structured evidence chain retrieval and pre-training large language model generation, large-scale generation of interpretable meteorological question and answer pairs in the vertical field is realized, and interpretability and credibility of question and answer results are improved.

The invention relates to a building energy calibration and prediction method based on a multi-task Gaussian process proxy model. The building energy calibration and prediction method is characterized by comprising the following steps: S1, data acquisition; s2, constructing an initial model; s3, sensitivity analysis and parameter screening; s4, sampling parameters; s5, performing batch simulation; s6, carrying out data sub-sampling; s7, constructing an agent model; s8, Bayesian parameter calibration is carried out; s9, predicting a physical model; s10, judging whether a preset standard is met or not; s11, correcting model deviation; and S12, final output is carried out. The method has the advantages that efficient and accurate calibration of building parameters and effective correction of model deviation are realized by constructing a weather decoupling agent model and a sequential calibration framework, and processing of high-dimensional data and construction of digital twinning can be supported.

This invention discloses a method and system for reusing discarded convolutional information in small-sample fine-grained image classification, belonging to the field of computer vision. The method includes the following steps: acquiring images of a support set and a query set, and extracting the backbone feature descriptors of the images using a convolutional neural network; after at least one convolutional block in the convolutional neural network, retrieving the discarded feature information of the convolutional block through a feature retrieval path to generate a retrieved feature descriptor; fusing the backbone feature descriptor with the retrieved feature descriptor to obtain a high-dimensional feature descriptor in a unified format; calculating the similarity between the images of the query set and the support set based on the high-dimensional feature descriptor, and performing small-sample fine-grained image classification based on the similarity, thereby realizing the reuse of discarded convolutional information. This invention can effectively retrieve discarded feature information in convolutional neural networks, improving the performance of small-sample fine-grained image classification.

This invention relates to the field of intelligent decision-making technology and discloses an intelligent decision-making system for agricultural irrigation based on artificial intelligence. The system constructs a collaborative architecture comprising a field perception layer, an edge computing layer, a cloud-based decision-making layer, and an irrigation execution layer, forming an irrigation task context through multi-source perception data. Based on this, a meta-reinforcement learning-driven multi-round agent-based search decision-making mechanism is introduced, combined with explicit self-reflection to achieve iterative optimization of the decision-making process. Simultaneously, a preference-aware reward evaluation model is constructed, achieving multi-objective trade-off modeling through latent variable mirror constraints and inverse autoregressive flow transformation, and globally optimizing the strategy using a cross-round benefit attribution mechanism to generate an irrigation scheduling scheme that meets water-saving, crop water requirement, and energy consumption constraints. This invention can improve the accuracy, stability, and adaptability of irrigation decisions.

The invention discloses a pointing object text positioning method based on iterative semantic visual association. The method comprises the following steps: acquiring a document image containing a pointing object and a natural language instruction; visual and language features are extracted, a pointing object bounding box is predicted, and a pointing mask is generated to extract pointing context features; the pointing context features and instruction semantics are spliced to generate a channel scaling coefficient, and space and channel modulation is performed on the visual features; using instruction semantics to generate FiLM parameters to modulate the enhanced visual features, and predicting an initial target bounding box through cross-modal attention fusion features; and starting an iterative refinement process, generating a target mask according to the current prediction frame and calculating a semantic-visual consistency score, modulating fusion features in combination with geometric difference and semantic information to predict a better bounding box, and outputting a result with the highest score after iteration is performed until a termination condition is met. According to the method, fine-grained spatial semantic understanding and high-precision iterative positioning capabilities are enhanced.

This invention provides a method and system for dynamic optimization of UAV flight trajectories based on neural networks, belonging to the field of data processing technology. The method includes: calculating the geometric moments of a two-dimensional contour; obtaining the central moment of the contour based on the geometric moments; normalizing the central moment and extracting Hu invariant moment features; combining shape fit features and Hu invariant moment features to form a digital environment feature vector; processing the digital environment feature vector with pose state information and a preset desired trajectory; fusing the results through a neural network model to obtain control parameter adjustment instructions; and, based on the control parameter adjustment instructions, real-time correcting of the UAV flight controller parameters to generate control signals driving the UAV actuators, thereby achieving dynamic tracking of the desired trajectory. This invention enables accurate dynamic tracking of complex trajectories by UAVs.

The application discloses a kind of real-time semantic segmentation method and system based on multi-resolution branch, it is related to graphics processing technical field.By obtaining image data, using depth multi-resolution network to carry out semantic segmentation to image data, obtain semantic segmentation feature map;Wherein, semantic segmentation process includes: to image data is carried out convolution feature extraction, obtains the spatial information and semantic information of image;Using different resolution branch network to the spatial information and semantic information of image respectively learns to obtain the feature map of each branch network;Fusion is the feature map of each branch network, obtains the final semantic segmentation feature map.The application extracts the semantic and spatial information of image from different resolution respectively, solves the problem of insufficient utilization of semantic and spatial information.The application also adds semantic enhancement module and feature fusion module, improve the effect of semantic information extraction while also solve the problem of spatial information being covered under multi-branch.

The application discloses a medical image segmentation method combining a double attention mechanism and a U-Net++, and related devices. The method comprises the following steps: acquiring a medical image and inputting the medical image into a segmentation network; extracting multi-scale coding features through an encoder, and generating multi-scale decoding features through a decoder; inputting the same-scale features into a double attention module, and sequentially executing channel attention and spatial attention; and finally, fusing the spatially weighted coding features to the decoder, and outputting a segmentation result. The application can be deployed in an artificial intelligence optimization operating system or an artificial intelligence middleware platform, high-efficiency calculation is realized by using a related function library, is suitable for computer audio-visual software, biological feature recognition software and other artificial intelligence application software, and the double attention mechanism and dense skip connection are combined, so that the feature expression capability is effectively enhanced, and the precision and robustness of medical image segmentation are improved.