204 results about "Boundary constraints" patented technology

The invention discloses an automatic river network grading and river topological relation identification method based on a graph theory, and relates to the technical field of hydrological geographic information. The method comprises the following steps: acquiring and cleaning a vector river network, a key point location and DEM data of a target drainage basin; constructing an initial river network graph model based on the line element connection relationship; integrating DEM topographic evidence and graph theory connection features, constructing and solving a global potential energy field equation containing topographic driving and boundary constraint, and calculating flow potential energy attributes of nodes of the whole network to determine a flow relationship; based on the flow direction relation, identifying topology abnormal structures such as strong connectivity components in the network, and performing ring breaking processing by using direction confidence to generate a ring-free directed network structure; and performing river grade division based on a topology transfer rule, and associating the key point location to a river network skeleton. According to the method, through global potential energy field solving and topological optimization, the problems that the flow direction of the plain micro-geomorphic area is difficult to recognize and complex loops cannot be graded are solved, and automatic construction of the river network topology is achieved.

The invention discloses a dynamic DEM spatial interpolation method which comprises the following steps: performing depression filling, flow direction analysis and confluence cumulant calculation on DEM data, and extracting a natural sub-basin unit by adopting a minimum catchment area threshold method; constructing a topographic feature matrix, performing refined second-level classification on the first-level drainage basin through an improved self-organizing mapping network, generating a hydrological response unit through boundary processing, and establishing a hydrological attribute library; fusing multi-source data, supplementing attribute interpolation such as underlying surface and rainfall, and constructing an interpolation auxiliary parameter system; a drainage basin is divided into regular grids as neurons, a dynamic neural network containing dynamic states and static attributes is constructed, and nonlinear mapping of DEM correction parameters is achieved through optimization of a dynamic activation function and a loss function. The method overcomes the defects that a traditional interpolation algorithm does not consider hydrological boundary constraints, a neural network model topological structure is fixed and the like, and the DEM interpolation precision and the hydrological simulation effect of the complex terrain area are improved.

The invention relates to the technical field of artificial intelligence, in particular to a sheet part performance rapid prediction method based on deep learning, which comprises the following steps: collecting multi-working condition simulation data to generate a training sample, constructing and coding a grid topological structure to extract multi-dimensional features, and inputting a perceptron to predict stress and evaluate errors after feature fusion and self-attention mechanism processing. According to the method, a structured training sample set is constructed by introducing simulation information, a geometric structure feature set is formed by combining node space coordinates, boundary constraints and a connection relation, so that mutual positions and constraint conditions among nodes are completely expressed in a graph structure, and through node-level feature extraction and feature fusion processing, a graph structure is obtained. According to the method, deep embedding of node geometric layout and boundary interrelation is realized, learnable expression of a stress evolution path in a space structure is established through local subgraph and context analysis, a multi-layer feature aggregation and attention mechanism is introduced in a node graph embedding process, and feature response expression of a key area is enhanced.

The invention relates to the technical field of tunnel engineering safety, in particular to a tunnel fire multi-stage damage collaborative prediction method and system, and the method comprises the steps: constructing a data simulation fusion model, and carrying out the processing of fire and tunnel structure parameter data, and obtaining a multi-working-condition parameter database; constructing a heat conduction equation, a fire plume and wall surface heat exchange relational expression, a boundary layer correction model and a comprehensive damage evaluation model, and analyzing the equation, the relational expression and the model by utilizing a multi-physical field coupling unified boundary condition to obtain a multi-physical boundary constraint condition; obtaining a multi-physical constraint loss function based on a network architecture and a multi-physical boundary constraint condition, and further obtaining a comprehensive constraint loss function of a temperature field-stress field-damage index; constructing a PINN-LSTM network architecture and a damage evolution analysis model based on the fire ending initial state coding vector, the fire course feature vector and the cooling process feature vector; and performing linkage analysis on the function, the frame and the model to predict and monitor the transient damage and the long-term creep damage of the tunnel.

The invention relates to an automatic guided vehicle (AGV) adaptive path tracking method based on MPC. The method comprises the following steps: S1, establishing a kinematic discretization error model based on the kinematic characteristics of the two-wheel differential AGV, processing a continuous kinematic equation by adopting an Euler discretization method, expressing a dynamic change relationship between a transverse deviation distance and an angle deviation in a state-space equation form, and generating a kinematic discrete state-space model; the method has the advantages that the continuous equation is processed by establishing the kinematics discretization error model and adopting the Euler discretization method, the deviation dynamic relation is expressed through the state space, the calculation process is simplified, the precision is kept, sensor data are fused, the wheel type odometer, IMU and laser data are integrated through the extended Kalman filtering algorithm, and the precision is improved. The real-time position and angle deviation are calculated, the positioning accuracy is improved, a model prediction controller objective function is designed, a weight matrix and boundary constraint are introduced according to constraint conditions, and the effect of obstacle avoidance constraint is combined.

The invention discloses a track optimization control method and system for a high-speed linear seven-axis manipulator, and relates to the technical field of data processing. The method comprises the following steps: acquiring a multi-manipulator grabbing task, dividing the task into single-manipulator grabbing sub-tasks, defining a starting point and an ending point for each sub-task, and generating a track boundary constraint set through a constraint hierarchical model; taking the termination point as a root node, generating a reverse extension tree in a feasible region range of the trajectory boundary constraint set through a potential field steering strategy, and outputting a multi-dimensional trajectory primitive; the multi-dimensional trajectory primitives are preprocessed, and a candidate trajectory set is screened out through a multi-objective optimization strategy; and integrating the optimized trajectories of the manipulators through a global trajectory conflict coordination mechanism, and detecting trajectory conflicts by adopting space-time alignment to generate a global optimal trajectory.

The invention discloses a high-formwork optimization design method and system based on BIM and finite element analysis, and the method comprises the steps: S100, building a three-dimensional BIM model of a main body structure, and creating an intelligent parameterization frame body family which is mapped with a main body structure family and has a self-verification function; s200, importing a parameterized family and linking the main body model, generating a plurality of groups of frame body arrangement schemes through spatial topology analysis, and performing a three-dimensional collision test after optimization; s300, setting dynamic boundary constraints and construction loads in stages, and generating a geometric and mechanical integrated data packet; s400, data packet parameters are automatically mapped to finite element software through a plug-in, models are built in stages, multi-working-condition analysis is carried out, if the parameters exceed the limit, frame body parameters are modified, iterative optimization is carried out until the parameters meet the standard, and a result is synchronized to a three-dimensional BIM model; and S500, a construction drawing is automatically generated, and visual disclosure and high formwork monitoring point selection are carried out. According to the method, design, computational analysis and construction application data linkage and closed-loop management from design to construction to monitoring are achieved, and the high formwork design efficiency and safety are improved.

The invention provides a coal ash composite material goaf closed filling parameter intelligent matching method, and belongs to the technical field of deep learning and mining engineering crossing. According to the method, dynamic optimization and accurate matching of the filling parameters are realized through combination of data driving and an intelligent algorithm. The method comprises four core links: multi-source data perception and fusion, a material performance prediction model, key parameter identification and boundary constraint, and intelligent matching and optimization decision. According to the method, the modeling capability of the model for the complex coupling relationship is improved through the attention mechanism and the feature cross network; the performance evolution trend of the material under different proportions and process conditions is accurately predicted, key regulation and control parameters are automatically identified, and efficient search and optimal matching of a parameter space are achieved through a strategy gradient method. According to the method, multi-source data can be fused, key parameters can be dynamically identified, and the intelligent matching and optimizing capability is achieved.

The invention discloses a method for generating a reference line in vehicle driving path planning and electronic equipment, and the method comprises the steps: obtaining initial data corresponding to an automatic driving vehicle, the initial data comprising driving map information, historical driving track information and vehicle parameter information; preprocessing the initial data to obtain target data; according to the target data, constructing boundary constraint information required by optimization based on the environment information; according to the initial reference line and the target boundary constraint information, generating target input information corresponding to a target reference line optimizer, and inputting the target input information to the target reference line optimizer, so that the target reference line optimizer performs reference line optimization processing according to the target input information to generate corresponding target output information, the target output information comprises a target reference line of the autonomous vehicle corresponding to a target driving road in the driving environment. Therefore, the reasonability and accuracy of the reference line can be improved.

The invention discloses a SiC MOSFET service life prediction method and device, equipment and a medium, and relates to the technical field of health detection of power electronic devices. The method comprises the following steps: acquiring an acquisition signal set acquired by a multi-channel sensor; the collected signals are preprocessed, smoothing and fusion estimation are carried out by adopting an extended Kalman filtering algorithm, and multi-dimensional degradation features including drain-source conduction voltage drift rate and thermal resistance change rate are extracted through feature engineering. And carrying out global optimization on the key hyper-parameters and the initial weight of the LSTM network by adopting an improved artificial fish swarm algorithm. The algorithm comprises an adaptive optimization strategy, a mixed local search operator and a boundary constraint processing mechanism. And training the optimized LSTM network by using the multi-dimensional degradation characteristics, establishing a mapping relationship between the multi-dimensional degradation characteristics and the device life, and obtaining a trained prediction model. And acquiring a to-be-predicted acquisition signal set to extract multi-dimensional degradation features and input the multi-dimensional degradation features into the prediction model to obtain a residual life prediction result.

The invention provides a trajectory planning method and device based on a space-time corridor and an automatic driving vehicle, relates to the technical field of computers, in particular to the fields of automatic driving, path and trajectory planning and the like, and can be used for trajectory planning, obstacle modeling and other application scenes. According to the specific implementation scheme, a space-time semantic corridor is constructed in a three-dimensional space-time coordinate system according to semantic information of a road reference line and a driving environment, and a motion track to be optimized is generated; matching a space-time corridor unit corresponding to the coordinates of the current track point from the space-time semantic corridor according to the to-be-optimized motion track, and constructing corridor boundary constraints based on the space-time corridor unit and the punishment cost; according to the road reference line and the motion trail to be optimized, constructing a trail following cost; constructing a trajectory optimization model according to the corridor boundary constraint, the trajectory following cost and a preset vehicle kinematics constraint; and solving the trajectory optimization model by using a hierarchical optimization algorithm to obtain a planned trajectory. According to the scheme, the trajectory planning speed and quality can be improved.

The invention relates to the technical field of underwater topographic surveying and mapping, in particular to a sea area water taking channel scanning survey method and system based on multi-beam measurement, and the method comprises the following steps: constructing a rigid support, fixing a depth finder through a three-way cable fastening device, and building a Boolean parameter model; sound velocity, tide level and original echo data are collected; performing space-time registration, coordinate conversion and filtering denoising on the data to generate corrected water depth; and constructing an irregular triangulation network by using hard boundary constraint, generating a digital elevation model, and extracting a siltation area. According to the invention, multi-beam wide sector scanning and multi-sensor attitude correction technologies are fused, a limited water area full-coverage measurement system is constructed, and the problems of terrain blind areas and detail loss are solved. A boundary constraint reconstruction algorithm is utilized to realize closed-loop processing from coordinate reference unification to micro-landform restoration, the accuracy of channel bottom pit capture and edge topography expression is improved, and data support is provided for sedimentation analysis.

The invention provides a geological envelope body automatic generation method. The method comprises the following steps: defining a target work area; performing sparse horizon picking of a top interface and a bottom interface on the geological envelope in the defined target work area to determine a top interface horizon, a bottom interface horizon and corresponding top interface and bottom interface data; based on the grid positions corresponding to all the pickup point positions of the top interface horizon and the bottom interface horizon, according to the range of the top interface data and the range of the bottom interface data, boundary lines of the geological envelope body at the top interface and the bottom interface in the target work area are determined; determining horizon depth values in all grid points of the top interface and the bottom interface; constructing a top interface boundary constraint and a bottom interface boundary constraint according to the horizon depth value and based on boundary lines of the geological envelope body at the top interface and the bottom interface to determine an intersection point between the top interface and the bottom interface of the envelope body so as to optimize the boundary lines of the geological envelope body at the top interface and the bottom interface; and generating a three-dimensional visualization model of the geological envelope based on the optimized boundary line.

The invention discloses an enhanced GAN extreme scene generation method and system considering weather factors, and relates to the technical field of scene generation in new energy power system planning and operation, and the method comprises the steps: extracting key factors from multi-time-delay meteorological data, forming an extreme label, converting the extreme label into a generation condition, and achieving the output through two-stage confrontation generation, wherein a scarce sample is stably amplified by WGAN-GP, a joint curve is generated according to conditions, and related structures, climbing and power boundary constraints are synchronously embedded to ensure statistical and physical consistency. According to the method, the authenticity of the distribution tail, the cross-variable correlation and the time structure can be considered when the sample is scarce. Extreme situation deduction of a novel power system can be supported while extreme situation coverage and credibility are improved, and an existing new energy scene analysis and planning process is integrated.

The invention relates to the technical field of computer vision, in particular to a sea surface image semantic segmentation method based on boundary feature constraint and electronic equipment, and the method comprises the following steps: S1, obtaining an image; s2, image preprocessing; s3, making a data set; s4, preprocessing the data set; s5, constructing a model; s6, training the model; and S7, outputting a result. According to the double-branch model based on the semantic branches and the boundary branches, expansion of the semantic information in the up-sampling process is restrained by extracting the boundary information and mutual query and constraint of the semantic information, the overall sensing value for a target is constructed in an auxiliary mode, and the problem of dynamic interference is effectively solved; by designing an attention multi-scale sensing module and a bidirectional attention fusion module, the segmentation precision of a water surface small target can be improved, and detail information can be supplemented; by providing the boundary query module and the boundary constraint module, a more accurate target boundary can be obtained on the premise of not introducing excessive parameters.

The invention discloses a multi-lane heterogeneous traffic joint modeling method and system based on a GNN-PINN coupling architecture. The method comprises the following steps: constructing a space-time dynamic graph coding lane changing rule through a GNN, embedding a traffic flow physical constraint through a PINN, and solving a source item; according to the method, CAV behaviors are described by using improved IDM in the mid-microscopic view, an LWR equation subjected to GNN lane change flow correction is solved by using PINN in the macro view, GNN and PINN closed-loop optimization is realized by combining a double-path back propagation mechanism, and a model is optimized based on a residual loss function fusing physical loss, data driving items and boundary constraints. The system comprises a dynamic graph topology module, a GNN coding module, a PINN solving module and the like, the multi-lane traffic dynamic description precision can be improved, the multi-scale characteristic is considered, real-time simulation is supported, and the system is suitable for heterogeneous traffic modeling.

The invention relates to the technical field of power distribution network toughness demand response, in particular to a power distribution network toughness improvement optimization method, system and device based on response prediction and a storage medium. Training an integrated decision tree model based on historical data to predict a response intention value of the cooling and heating load user, and determining a temperature regulation boundary and a response frequency upper limit according to the response intention value; generating an initial scene set through source load uncertainty sampling, extracting a typical scene by adopting a transportation distance scene reduction method, and introducing an overall offset constraint and a single-point extreme value constraint to perform two-dimensional limitation on scene probability distribution; a scene probability combination enabling the load recovery value to be minimum is searched in a probability distribution domain, and a scheduling scheme enabling the key load recovery value to be maximum is solved under the condition that the power flow constraint, the cold and heat power balance constraint and the temperature regulation boundary constraint are met; and carrying out probability weighting on the load recovery values of different scenes to obtain a toughness evaluation value, and carrying out sensitivity analysis.

The invention provides an artificial intelligence-based city scene generation method and a computer readable storage medium, and the method comprises the steps: firstly, carrying out the automatic recognition and extraction of different functional regions based on input graphic data, and carrying out the vectorization of the different functional regions into polygons with structural features; matching a three-dimensional model from a model library according to polygon features and generating an initial layout containing positions and orientations; constructing a differentiable boundary constraint function for each functional region to define a legal region; then defining a differentiable objective function including a boundary loss item and a collision loss item; and finally, by taking the initial layout as a starting point, optimizing an objective function through a gradient descent method, and iteratively adjusting the layout of each three-dimensional model to obtain a final city scene. According to the method, the spatial layout rule is converted into the computable loss function through the differentiable optimization framework, the automatic and intelligent generation of the city scene is realized, the layout generation efficiency is effectively improved, and the layout rationality is ensured.

The invention discloses a water engineering scheduling intelligent parallel caching system for intelligent water conservancy and an optimization method. The system comprises a constraint-driven asynchronous hierarchical caching module, a topology-sensitive cumulative hash module and a polymorphic hierarchical storage module. According to the system, a three-level framework of thread local buffering, global batch processing and an generational sliding window is adopted, and boundary constraint pre-verification is executed on a scheduling scheme; decision vectors are rearranged according to the upstream and downstream topological relation of the water engineering, an upstream engineering decision preferentially participates in Hash calculation and is endowed with a higher weight, and the weight is decreased progressively according to an exponential function; target value arrays are dynamically distributed, and volume class, flow class and time sequence class constraint violation degrees are stored in a layered mode; according to the method, through a topology-aware Hash mechanism and a self-adaptive caching strategy, the recognition precision and the caching hit rate of similar schemes are remarkably improved, repeated calling of a hydraulic simulator is effectively reduced, and the calculation efficiency of water engineering scheduling optimization is improved.

The invention provides an intelligent interpretation method for a dual-temporal remote sensing image, which is characterized in that a boundary constraint change detection model BCnet based on a visual basic model is constructed by the method, the universal semantic representation potential of the visual basic model is fully excavated, and a difference detail enhancement module and a multi-scale edge enhancement module are introduced, so that the visual basic model can be fully interpreted. The problem that high-frequency information is lost in the direct migration process of the visual basic model is solved, and fine description of tiny changes and complex textures is achieved. On the basis, an edge feature constraint strategy is designed, an edge feature aggregation module is combined, boundary supervision signals are introduced into a feature space and a prediction space at the same time, and the continuity and geometric integrity of a change detection result in the space structure are enhanced. Experimental results prove that the BCnet has the advantages of high robustness and precision in high-difficulty scenes such as dense building groups and complex edges.

The invention provides a predictive intelligent capacity expansion and contraction method for a large language model inference service and related equipment. The method comprises the following steps: calculating a first instance demand quantity based on a GPU cache utilization rate of a corresponding historical time period in historical monitoring data; obtaining a real-time GPU cache utilization rate of a large language model inference service running in the computing cluster, and calculating according to a preset second target utilization rate to obtain a second instance demand quantity; fusing the first instance demand quantity and the second instance demand quantity to obtain a capacity expansion and contraction prediction instance quantity; boundary constraint is carried out on the capacity expansion and shrinkage prediction instance number so as to be limited between a preset minimum instance number and a preset maximum instance number, and a target instance number is obtained; and executing capacity expansion and contraction operation based on a comparison result of the target instance number and the current running instance number. According to the method, a two-way prediction fusion mechanism of historical monitoring data and real-time monitoring data is introduced, and the response speed and stability of large language model reasoning service are remarkably improved.

The invention discloses a feature-preserved curvature-adaptive triangular mesh generation method. The method comprises the following steps: acquiring a geometric adaptive size field according to a curved surface set based on a geometric size field construction strategy; based on the geometric adaptive size field, performing curve discretization processing on the curve set according to the geometric adaptive size field to obtain a discrete curve grid with reserved curve features; taking the discrete curve grid as a rigid boundary constraint condition, and performing curved surface discretization processing on the parameterized curved surface patches according to a geometric adaptive size field through a leading edge propulsion method and a triangulation coupling strategy to obtain discrete curved surface grids of all the parameterized curved surface patches; and carrying out local grid merging and optimization on the discrete curved surface grid to obtain a uniform curved surface grid with geometric continuity, namely a curvature self-adaptive triangular grid with reserved features. According to the method, the technical problems that the quality of the boundary grid cannot be ensured, the efficiency is low and the convergence is poor in the existing method are solved.

The invention provides a multi-source domain invariant acoustic feature extraction method and system for an equipment operation state, and belongs to the technical field of equipment maintenance, and the method comprises the steps: constructing a multi-source domain invariant acoustic feature extraction network based on a DANN model, and the network comprises a feature extractor, a classifier, a domain discriminator and a multi-domain acoustic feature class boundary constraint module; the feature extractor extracts high-dimensional features of sound signals, the classifier carries out fault mode recognition, and the domain discriminator realizes domain prediction. The multi-domain constraint module generates an embedding space, and calculates the maximum mean value difference, the local maximum mean value difference and the Euclidean distance among different source domain features. The network constructs a loss function by taking minimization of inter-domain difference and classification loss and maximization of inter-domain distance and domain discrimination loss as targets, and carries out adversarial training through multi-source tagged acoustic data. According to the method, domain invariant features are extracted by using adversarial learning and hidden space constraint alignment of multi-source domain acoustic features, so that the influence of feature offset under a cross-working-condition condition is effectively reduced, and the fault mode recognition accuracy is improved.

The invention discloses a task allocation method and system based on an improved whale optimization algorithm framework, relates to the technical field, and is used for optimizing order type adaptive cross-domain traffic control network task allocation and improving the key task response capability of a time-sensitive traffic system. ICWOA initializes a population through Chebyshev mapping, and introduces Levy flight disturbance to enhance the optimization ability; dynamically balancing local and global search by means of adaptive parameters; relieving population diversity attenuation through randomness retention, diversity maintenance and boundary constraint; dimension type pinhole imaging reverse learning is fused to reduce high-dimensional optimization dimension interference. According to the algorithm, the convergence speed and the solving precision are better, sub-second calculation time is kept under different task scales, the task distribution efficiency is improved, and the time-sensitive scene task success rate is remarkably improved. The method solves the problems that an existing algorithm is insufficient in adaptive order type'order application-order sending 'structure and prone to falling into local optimum, population diversity attenuation and calculation speed.

The invention discloses an agricultural robot trajectory planning method and system with optimal energy consumption under collision constraint. The method comprises the following steps: acquiring historical state sequence data and environmental map information of a target machine and peripheral machines in real time by a self-machine; identifying the operation intention of the target machine, and outputting a multi-modal prediction trajectory; the method comprises the following steps of: generating a global coarse reference track covering an operation area by taking completion of an own operation task as a main target; generating a space-time safety corridor by adopting a space-time inheritance strategy, and reserving a safety space for the dynamic obstacle; constructing an optimization problem cost function, and comprehensively considering reference trajectory cost, collision risk cost and energy consumption cost; and under the boundary constraint of the safety corridor and the kinematics constraint of the robot, solving the optimization problem by using a numerical optimization solver to obtain an optimized local trajectory. According to the method, safe, efficient and energy-saving collaborative operation can be realized by fusing multi-modal trajectory prediction of the dynamic obstacle, refining energy consumption modeling and risk perception planning.

The invention discloses an AC / DC system load flow calculation method and device based on trajectory uniformity, and belongs to the technical field of power system calculation and analysis. The method comprises the following steps: inputting topological parameters of an AC / DC system, node power injection data and boundary constraints of variables; according to the topological parameters and node power injection data, an AC / DC system feasible power flow solution model containing voltage constraints is constructed, and the voltage constraints are converted into equality constraints through slack variables; constructing a QGS based on a feasible power flow solution model, wherein a stable equilibrium point of the QGS corresponds to a solution or a least square solution of a power flow equation; tracking a QGS system track by adopting a numerical integration method, and carrying out integration from an initial point until a convergence condition is met; and judging whether the QGS system converges to a feasible power flow solution or a least square solution according to an integral result, and outputting a corresponding result. According to the method, any initial point in the power flow solution neighborhood can be reliably converged to the power flow solution, and the robustness and practicability of power flow calculation are remarkably improved.

The invention discloses a territorial space current situation evaluation model method fusing multi-source heterogeneous data, which comprises the following steps of: generating grid units based on topological boundary constraints of an administrative region, and establishing a multi-level index of the grid units of the administrative region through STree; integrating the topological boundary constraint and a point-grid mapping technology to realize that vector space file elements and a table generate a spatial index relationship in the grid unit; according to the grid units and the spatial index relationship, performing label index calculation in the grid by utilizing a labeling calculation tool with natural resource business semantics and fuzzy spatial semantics to obtain a grid label table; and calculating a label index weight based on a CRITIC algorithm fused with the spatial weight matrix, and obtaining a natural resource comprehensive evaluation according to the label index and the label index weight. According to the method, multiple space collaborative evaluation according to the industry field can be supported, and the space calculation scale is effectively unified.

The invention discloses a nonlinear system dynamic boundary constraint particle filtering method, and belongs to the technical field of deep space exploration. The implementation method comprises the following steps: independently sampling all nodes of the flexible lander, and introducing constraints in a particle filtering initial sampling link to enable initial particles to meet the constraints; in the filtering process, as many particles as possible fall within the constraint range in the initial stage of filtering by widening the constraint boundary, and the constraint boundary is continuously shrunk to be gradually close to the real constraint range, so that the situation that a large number of particles fall out of the constraint boundary due to the fact that the real constraint range is too narrow is avoided. According to the method, the constraint boundary is dynamically adjusted, the number of effective particles is increased, and a higher-precision multi-node state estimation result of the flexible lander is obtained; according to the multi-node state of the flexible lander, the particle diversity is ensured through dynamic scaling constraint boundaries, and filtering stop caused by particle degradation is prevented; and realizing real-time trajectory planning and guidance control of the flexible lander by using a node state estimation result.

The invention discloses a track video inspection method and system based on deep learning, and relates to the technical field of video inspection, and the method comprises the steps: obtaining map information, analyzing the effective shooting width of unmanned aerial vehicle inspection, defining track boundary constraint, and analyzing the effective width of track shooting, so as to determine different flight path paths; the fitness value of the flight path is evaluated to determine the optimal path, the turning energy consumption and the stable flight energy consumption of the optimal path are analyzed, and step-by-step coordinate points are added to determine the track segment of the inspection task. According to the method, the segmented task set is further optimized and combined with the real-time energy consumption to form the track segmented task set, so that the effect of converting a long path into a small-scale inspection task of a plurality of sub-paths is achieved; it is guaranteed that the track collection task can still be efficiently executed under the condition that the illumination condition is insufficient, and definition of the weak light collection task set provides clear guidance for follow-up height adjustment and infrared light supplement optimization.

This invention relates to a multimodal visual understanding method based on consistency learning and hybrid feature extraction. It includes constructing an end-to-end fine-grained consistency learning framework, introducing a hybrid region extractor, fusing local details and global semantics to generate high-quality hybrid visual cue embeddings, combining self-reconstruction loss and latent spatial consistency loss to force the model to establish explicit alignment between the input visual cue and the output segmentation label, utilizing the geometric boundary constraints of the localization task for description generation, and simultaneously optimizing localization accuracy using the semantic depth of the description task. Furthermore, it constructs a detailed localization index expression and segmentation task to enhance the model's reasoning ability for complex long text instructions. The aim is to address the problems of feature fragmentation and insufficient accuracy in existing large models for fine-grained visual localization and description tasks. Compared with existing technologies, this invention has advantages such as high accuracy and strong generalization ability in pixel-level localization and fine-grained description.