144 results about "Local topology" patented technology

The invention belongs to the technical field of electric power automation, and discloses a power grid load flow calculation method based on a graph neural network and a related device. The method comprises the following steps: acquiring power grid data to generate a power flow data set; constructing a node feature vector and an edge feature vector based on the power flow data set; splicing the node feature vectors and the position codes of the nodes in the power grid to obtain splicing features; and inputting the splicing feature and the edge feature vector into a pre-trained graph neural network model to obtain a power grid load flow calculation result based on the graph neural network. Compared with a traditional Newton-Raphson method and a graph neural network method considering a local topological relation, the method has the advantages that the calculation speed can be remarkably increased, and meanwhile, the accuracy and reliability of a result can be kept; the method has the advantages of being high in adaptability, wide in application scene and the like, can adapt to power grids of different scales and topological structures, has good generalization ability, and can be applied to multiple application scenes such as real-time power flow calculation and future state power flow prediction.

The invention belongs to the related technical field of brain image processing, and provides an Alzheimer's disease classification method and system based on topology perception and a group hypergraph in order to solve the problem of inaccurate classification of the Alzheimer's disease in the prior art. Constructing a dynamic function connection network sequence through a sliding window strategy; a local topology perception encoder and a global topology perception encoder are respectively used for extracting local topology features and global topology features of each time window, deep interaction and fusion are carried out, and comprehensive feature representation of a tested level is generated; according to the method, each subject is used as a hypergraph node, hyperedges are constructed on the basis of comprehensive feature representation of a subject level and by combining feature similarity calculated by diffusion tensor imaging features and clinical embedded features of the subject, then a group hypergraph is constructed, a classification result is obtained by using a hypergraph neural network, and the early classification diagnosis accuracy of the Alzheimer's disease is effectively improved.

The invention discloses a three-dimensional physical field real-time prediction method and system based on geometric deep learning and physical constraint, and relates to the technical field of computer-aided engineering and artificial intelligence. The method comprises the following steps: directly extracting native boundary representation data (B-Rep) of a three-dimensional model from a computer aided design system; constructing a heterogeneous dual graph taking a parameterized curved surface as a graph node, skipping finite element grid division, and aggregating local and global topological features by using a graph neural network; in combination with a physical information driving mechanism, a partial differential equation (PDE) residual error is introduced as a loss function for constraint training, and generalization prediction of a novel geometric structure is realized; and finally, the physical field state quantity is predicted through direct regression and is rendered in real time. An incremental reasoning mechanism based on a local topology subgraph is adopted, millisecond-level physical field real-time feedback under design modification is achieved, and the method is suitable for scheme rapid screening and trend prediction in the initial stage of design.

The invention belongs to the technical field of space intelligence, and relates to a dynamic obstacle intervention-oriented space intelligent scene self-reconstruction navigation system, which comprises a dynamic intervention sensing module, a dynamic obstacle intervention processing module, a dynamic obstacle intervention processing module, a dynamic obstacle intervention processing module, a dynamic obstacle intervention processing module and a dynamic obstacle intervention processing module, the spatial semantic topology reconstruction module is used for automatically updating a spatial semantic map and a topology connection relation according to the dynamic intervention event set; the local topology rapid recombination module is used for carrying out rapid recombination on the intervened and influenced local topology based on a topology incremental learning principle; and the perception and navigation bidirectional consistency optimization module is used for establishing a reverse feedback path between perception and navigation. According to the invention, the real-time response and structure-level self-adaption of the navigation system to environment change can be realized, so that the stability and consistency of navigation decisions can be maintained in changeable scenes. The invention further provides a space intelligent scene self-reconstruction navigation method oriented to dynamic obstacle intervention.

The invention relates to the technical field of knowledge fusion, and discloses an interactive health science popularization question and answer guidance system for constructing a medical knowledge graph, comprising: a guidance instruction generation module extracting attribute variation dimensions between to-be-aligned data and candidate nodes and generating a guidance instruction; the feedback analysis module converts the feedback data into a feedback correction vector; the atlas fusion actuator determines an initial fusion weight according to the coordinate distance, determines a weight correction coefficient based on the topological centrality of the target node, constructs a weighted feature update function by using a feedback correction vector and a local topological relation to calculate a feature increment, and completes feature convergence fusion; according to the method, a priori constraint force balance feedback feature pointing value provided by a map structure is utilized, it is ensured that heterogeneous knowledge fusion conforms to medical ontology logic, and the global stability of science popularization map evolution is guaranteed.

The invention provides a communication optimization method for topological table persistent storage in efficient parallel computing, belongs to the technical field of storage communication, and aims to avoid pseudo sharing by aligning memory allocation through cache lines, expand local topological coverage through a topological entropy increment driven prefetching mechanism, and improve the reliability of the topological table persistent storage. Boundary processing is accelerated through a pre-calculation period mapping lookup table and a frequency domain transfer function vector, synchronization overhead is optimized through a concurrency control mechanism perceived by a read-write ratio, targeted cache preloading is achieved through stability and jitter degree two-dimensional evaluation, bandwidth consumption is reduced through an increment synchronization mechanism, and the stability of the cache is improved. The communication template is selected or the communication parameters are generated through adaptive conversion by matching the matching degree decision, and the technical problem that the parallel computing performance is reduced due to the fact that the communication overhead is too large in the topological table persistent storage process is solved.

The invention discloses a dexterous hand environment article sensing and modeling algorithm based on body sensing. The algorithm comprises the steps that a mechanical arm drives a dexterous hand to approach an object; tactile-spatial data acquisition: a three-axis force sensor is arranged at the tail end of each finger of the dexterous hand, and the system is in contact with an object through multiple fingers of the dexterous hand to acquire multi-dimensional information of the appearance, surface characteristics and local topology of the object; active exploration sampling, multi-finger movement of the mechanical arm and the dexterous hand, and active penetration of unknown object surface and enhanced sampling of local features through the dexterous hand by the system; the method comprises the following steps: modeling an object attribute, analyzing a mechanical response sequence obtained by tactile-spatial data acquisition and active exploration sampling, modeling a target object in three dimensions of elasticity, hardness and local geometric morphology, and realizing accurate identification and modeling of an unknown object appearance form and an object attribute in an environment. The distribution uniformity of the tactile point cloud and the model integrity are improved, the hardware dependency is reduced, and the environmental adaptability is enhanced.

The invention belongs to the technical field of image recognition and power grid capital construction sites, and particularly relates to a power grid capital construction dynamic risk recognition method based on a multi-modal image and LLM deduction, and the method comprises the steps: limiting the relation between an inspection object and a design based on a stress chain and an equipotential chain, and generating a constraint framework; performing multi-channel sampling on objects in the constraint framework; after the channels are aligned, a state evidence packet is generated, and the design relation and the state evidence packet are aligned; a stress chain and an equipotential chain are assembled, and a local topological graph is obtained through cutting and minimum necessary path selection; based on the local topological graph, mapping mechanism elements of the nodes and the edges into a fixed semantic slot position and action mode template, and generating mechanism description; and outputting a dynamic risk source list based on the mechanism description and the local topological graph. According to the method, the multi-modal image is subjected to constraint mapping to form the mechanism elements and the minimum necessary topology, and deduction is carried out on the stress chain and the equipotential chain, so that the dynamic risk source is accurately identified, and evidence-level traceable decision output is realized.

The invention relates to the technical field of retail management, in particular to a retail end online management method and system based on concurrency control, and the method comprises the steps: constructing a retail knowledge graph based on historical transaction logs through employing a space-time interaction entropy weighting algorithm; calculating a local topology edge density increment by applying a streaming dense subgraph increment mining algorithm, and determining a topology aggregation state of a target node; if the state is the mixed concurrent state, constructing an inventory time slice mirror image containing a reference inventory domain and a dynamic increment domain; and writing the request into the lock-free annular buffer during checking, and merging data by adopting a time sequence alignment algorithm after checking to carry out atomized updating. According to the method, static inventory and dynamic sales are physically isolated by constructing the inventory time slice mirroring and lock-free buffer mechanism, the read-write conflict in the scene of inventory while sales is solved, and zero-blocking hybrid concurrence is realized while the final consistency of inventory data is guaranteed.

The invention relates to the field of image recognition, discloses an image recognition method and an image recognition system, and systematically solves the core pain point of traditional document recognition through optical-topology fusion processing and a dynamic resource allocation mechanism. The image recognition method is composed of an acquisition module, a grid module, a phase module, a setting module and a distribution module, pixel brightness is calculated and coded based on document RGB data, and a two-dimensional coding matrix is generated; constructing a geometric correction grid, forming a nonlinear constraint field, and enhancing the anti-deformation capability of the image; generating spiral phase light waves in the constraint field by using a spatial light modulator, and generating a time-varying phase map; determining a local topology index by detecting the number of phase jump times; and extracting the closed boundary region as a character block, and generating an analysis result file. The system breaks through traditional limitation, improves image geometric correction precision, feature extraction sensitivity and boundary judgment accuracy, efficiently completes document analysis, and is suitable for scenes such as document digitization and information retrieval.

The invention belongs to the technical field of computer vision and smart agriculture, and discloses a leaf point cloud reconstruction method and system based on projection constraint and hybrid supervision, and the method comprises the steps: obtaining multi-modal data of seedlings, and carrying out the preprocessing of the multi-modal data; constructing a point cloud self-encoding network with feature circulation consistency to output a prediction point cloud; the method comprises the following steps: according to the dynamic three-dimensional geometric loss of a sample integrity label dynamic switching constraint mode, using a micro rendering technology to project a predicted point cloud and then calculate the predicted point cloud and a two-dimensional segmentation mask so as to calibrate the micro projection consistency loss of an edge; the local topology regularization loss of complementing the smoothness of the region is ensured; and dynamically adjusting the weight of each loss item by adopting a course learning strategy to carry out network training. According to the method, sub-millimeter-level edge reconstruction is realized by using two-dimensional image information which is easy to obtain, and the quality of incomplete blade point cloud completion and the accuracy of blade area measurement are remarkably improved.

The invention belongs to the technical field of power distribution equipment monitoring, and relates to an intelligent power distribution equipment monitoring and fault early warning system and method, and the system comprises an acquisition unit which is used for obtaining an electric parameter, an environment parameter and an operation state quantity of power distribution equipment and forming time synchronization feature data; the topology unit is used for generating a local topology model according to the relation between the cabinet, the loop and the busbar and outputting topology attributes; the evaluation unit is used for generating an operation state evaluation result based on the feature data and the topological attributes and updating an early warning threshold value; the governance unit is used for performing deduplication, suppression and credibility calculation on the alarm to form a disposal level; the interlocking control unit is used for checking and outputting a control action when the disposal level meets an interlocking condition; and the operation and maintenance unit is used for generating a preventive test plan and updating parameters based on a test result. According to the technical scheme of the invention, topological association and threshold update are introduced into monitoring, evaluation, disposal and control links, so that state recognition and control execution are kept consistent.

The invention relates to the technical field of artificial intelligence, in particular to a knowledge distillation acceleration method and system for a multi-modal large model, and the method comprises the steps: obtaining multi-modal input data; feature mapping of topology maintenance is executed, topological representation of a multi-modal feature space is constructed, and a feature mapping module is designed based on a residual module; a topological hierarchical attention mechanism is applied, and multi-head topological attention fusion is realized through sensing of a local topological attention unit and a global topological structure; topology-driven multi-task knowledge distillation is executed, five cross-modal distillation tasks are designed, and distillation loss is calculated based on topological similarity measurement; a topology-aware model acceleration technology is adopted, adaptive quantization is performed based on feature topology importance, neural network structure search for topology maintenance is executed, a knowledge distillation process is guided through a topology principle, a topological structure relationship of a feature space is maintained, and multi-modal knowledge is effectively transmitted. Experiments show that compared with an original large model, the method has the advantage that the size of the model is reduced by 80-90%.

The invention discloses a power distribution network abnormity positioning and warning method based on a time sequence power flow model, and relates to the technical field of power distribution network power flow monitoring. Layered verification is carried out on source diameter nodes of the power distribution network, power flow parameter verification is carried out, a real-time power flow graph under a dynamic time sequence is generated based on a constructed time sequence power flow model so as to carry out dynamic time sequence feature extraction, and then comparison is carried out based on an extraction result and a power flow parameter corresponding to a power flow starting point time sequence so as to obtain a dynamic time sequence feature extraction result. Meanwhile, anomaly detection is carried out, an anomaly detection dynamic tidal current diagram is obtained, the situation that tidal current parameters of a plurality of source path nodes in a certain hierarchy are abnormal at the same time is positioned, finally, hierarchical positioning and local topology adjustment of the abnormal source path nodes are carried out, and the abnormal source path nodes in the power distribution network are isolated. Meanwhile, warning scheme verification and dynamic power flow graph updating are carried out in combination with a time sequence power flow model, and the problem that in the prior art, the false warning rate is high when the topological structure changes in the topological correlation operation process of the power distribution network is solved.

The invention provides a molecular property prediction method and system, computer equipment and a storage medium, and belongs to the field of compound intelligent analysis, and the method comprises the steps: obtaining an SMILES character string of a to-be-recognized molecule; converting the character string to obtain a molecular object, extracting a fingerprint vector from the molecular object, and performing linear transformation processing to obtain chemical substructure features; constructing a two-dimensional molecular map based on the molecular object, and extracting features of the two-dimensional molecular map to obtain local topological structure features; rendering the molecular object to generate a molecular image and extracting image features; performing feature splicing on the chemical substructure features, the local topological structure features and the image features to obtain fusion features; mapping the fusion feature into an output space to determine a predicted value, and generating a molecular property prediction result of the to-be-identified molecule based on the predicted value. According to the method, complementation among different information types can be realized, more information of molecules can be explored from different angles, and the accuracy and the universality of predicting molecular properties are enhanced.

The invention relates to the technical field of track obstacle avoidance, and discloses a track robot path obstacle avoidance data updating method and system, and the method comprises the steps: carrying out the vector assembly of multi-dimensional features in a real-time environment data flow of a track robot, and obtaining a primary feature vector, carrying out abnormal mode identification on the primary feature vector and a historical normal mode library to obtain a dynamic abnormal feature; according to the dynamic abnormal features, reconstructing a local topology of a pre-stored path knowledge graph to generate an obstacle influence domain; analyzing the field intensity distribution change trend of the obstacle influence domain to obtain a gradient field, and mapping the dominant direction of the gradient field into a conflict resolution strategy; correcting a predetermined path of the orbital robot to obtain a conflict-free path trajectory fragment; performing efficiency judgment on the conflict-free path trajectory fragments to obtain a multi-dimensional evaluation result; fusing multi-dimensional evaluation results to obtain an emergence type motion strategy; the efficiency of updating the path obstacle avoidance data of the rail robot can be improved.

The invention discloses a multi-source geographic vector data matching method based on topological relevance, which comprises the following steps: collecting multi-source basic geographic vector data to form a preprocessed data set; constructing multi-index data quality evaluation based on local topological features; extracting and classifying spatial topological relation edges; constructing a weighted adjacent matrix of the target map; a neighborhood topology consistency measurement index is calculated, and neighborhood topology consistency quantization and adaptive weight fusion are carried out; establishing a global optimization objective function by taking the confidence coefficient matrix as input, solving a maximum weight bipartite graph matching problem with topology constraints by taking the topology retention rate as a hard constraint condition, and outputting a global optimal matching scheme; and constructing a quantitative evaluation index topology deviation measure based on topology invariant difference, judging whether a matching scheme has a significant topology conflict, and performing hierarchical repair according to the conflict condition. According to the method, the reliability and logic self-consistency of multi-source vector data fusion are improved, and the key problem of topological contradiction in the prior art is solved.

The invention discloses a material granularity identification method and system based on appearance analysis, and the system generates a standardized image matrix through image collection and preprocessing, employs a two-channel neural network structure to extract spatial positioning features and morphological structure features, carries out the fusion of the features to form a guide mask pattern, and carries out the recognition of the granularity of a material based on an improved active contour evolution model. Position constraint and form constraint are introduced into an energy function at the same time, dynamic evolution of the contour is achieved, the system monitors boundary consistency in the evolution process, and local topology reinitialization is triggered when the boundary consistency is lower than a threshold value so as to guarantee segmentation stability. After evolution is completed, a final contour area is extracted, particle size distribution data are calculated in combination with image scale parameters, and a particle size recognition result is output, automatic recognition and statistical analysis of the material particle size are achieved, and the method is suitable for particle material detection and distribution evaluation scenes.

The invention discloses a blade structure generation method based on bearing capacity flow topology optimization, and relates to the technical field of wind power generation equipment. The method comprises the following steps: establishing a three-dimensional finite element reference model of a wind turbine blade, and identifying and intercepting a local design domain containing a web-beam cap connection area from the model; finite element modal analysis is carried out on the blade or the segment of the blade, and one or more vibration node lines are extracted from the vibration mode of the blade or the segment of the blade; in the local design domain, establishing and solving a local topological optimization mathematical model which takes minimization of material interface peak stress as an optimization target and takes a dynamic geometric constraint which requires optimization and generated material distribution to follow a vibration node line in space as a core constraint condition; and generating a geometric model of the blade connection structure according to an optimization solution result. According to the method, the static reliability and the dynamic anti-fatigue performance of the key connection area of the blade can be prospectively and integrally improved.

The invention provides a point-to-point communication method and system based on self-organizing direct connection, and the method comprises the steps: obtaining network state information in a plurality of time periods, obtaining a historical communication data set sequence, carrying out the construction of a communication topological graph, and carrying out the time sequence feature modeling, and obtaining a dynamic communication time-space diagram sequence; performing space-time correlation feature analysis and time sequence communication state evolution on the dynamic communication space-time diagram sequence to obtain an expected communication state data set corresponding to each communication link in a future period; determining a communication device group needing point-to-point communication, and performing multi-constraint adaptive path decision on a local topology link of the communication device group according to the expected communication data to obtain a standard communication link; and performing signal transmission among the communication equipment groups by using a standard communication link to obtain a received signal, performing time-frequency transformation, image denoising and feature enhancement on the received signal to obtain an enhanced frequency domain feature map, and performing inverse transformation on the enhanced frequency domain feature map to obtain an enhanced received signal.

The application belongs to the field of underwater acoustic routing protocol, and relates to a kind of underwater acoustic communication network routing method of dynamic optimization of communication resource.The underwater acoustic communication network includes a source node, a destination node and multiple relay nodes, and the method includes the following steps: step 1) the source node initiates the initialization of the underwater acoustic communication network, and each relay node constructs a local topology table through the exchange of metadata packets;step 2) the sending node determines a forwarding candidate set according to the position information and residual energy of the neighbor nodes;step 3) the sending node selects the relay node with the highest forwarding priority in the forwarding candidate set as the optimal relay;step 4) the sending node performs communication resource optimization to obtain communication parameter configuration;step 5) the sending node sends data packets to the optimal relay node according to the configured communication parameters;step 6) it is judged whether the receiving node is the destination node: if it is the destination node, the process is ended;if not, steps 2) to 5) are repeated until the receiving node is the destination node.

The present application belongs to the technical field of data processing, and discloses a multi-modal project data analysis method, which comprises the following steps: collecting heterogeneous modal project data; based on adaptive multi-scale filtering technology, denoising and timestamp correction are performed on each modal project data, and cross-modal time alignment is performed, and a multi-modal signal is output; a multi-level Boltzmann machine energy network is constructed, and a joint probability distribution of the multi-modal signal is modeled; a dynamic complexity measurement mechanism based on topological entropy is designed, which is used for representing project structure complexity, and a topological entropy dynamic field is generated; the topological entropy dynamic field is used as a constraint condition, a multi-dimensional covariant field framework is used, nonlinear coupling and space-time propagation among the multi-modal signals are simulated, the multi-modal signals are abstracted into string vibration modes, local topological defects in the topological entropy dynamic field are identified, and an abnormal fluctuation atlas is generated; and the analysis of complex multi-modal data in the project running process is more intelligent and controllable.

The invention discloses a knowledge graph fusion method, device and equipment based on label propagation. The method comprises the following steps: acquiring a label propagation command, wherein the command comprises a knowledge graph list and a result output path; respectively executing a label propagation algorithm on each knowledge graph in the knowledge graph list, and determining initial label value distribution of each knowledge graph; determining an equivalent entity set among different knowledge maps; based on an initial label value and a local topological structure of each equivalent entity in the equivalent entity set in the knowledge graph, fusing the initial label value and the topological structure to obtain a fused entity, a fused label value and a fused topological structure; and according to the fusion entity, the fusion label value and the fusion topological structure, fusing the at least two knowledge maps in the knowledge map list to obtain a fusion result, and storing the fusion result in a result output path. According to the embodiment of the invention, cross-graph complex semantic calculation and conjoint analysis are supported.

The present application relates to the technical field of track obstacle avoidance, and discloses a track robot path obstacle avoidance data updating method and system, the method comprising: assembling a plurality of dimensional features in real-time environment data flow of the track robot into a vector group to obtain a primary feature vector, and performing abnormal mode identification on the primary feature vector and a historical normal mode library to obtain dynamic abnormal features; reconstructing a local topology of a pre-stored path knowledge graph according to the dynamic abnormal features to generate an obstacle influence domain; analyzing a field strength distribution trend of the obstacle influence domain to obtain a gradient field, and mapping a dominant direction of the gradient field into a conflict resolution strategy; correcting a predetermined path of the track robot to obtain a conflict-free path trajectory segment; performing efficiency determination on the conflict-free path trajectory segment to obtain a multi-dimensional evaluation result; and fusing the multi-dimensional evaluation result to obtain an emergent motion strategy; the present application can improve the efficiency of track robot path obstacle avoidance data updating.

This invention relates to the field of computer graphics processing and 3D modeling technology, specifically to an automated intelligent generation system for 3D modeling of shoe exteriors. The system includes: data acquisition and streaming, where an image acquisition terminal generates and sends four-channel image data using a high-precision sensor array; feature reconstruction and representation, where a topological semantic reconstruction server extracts semantic features and constructs an initial geometric representation using an implicit function based on pixel alignment and a decoder to predict the symbolic distance function; topology optimization and segmentation, where topological streamline optimization and component semantic segmentation are performed based on the curvature field to generate a structured 3D model with independent component structures; and interaction and local updates, where an interactive terminal responds to user editing operations on independent component structures, sending local topology update commands, and the server performs parametric recalculation of the local mesh based on the commands. This invention solves the problem of the separation between geometric structure and semantic information, achieving real-time and refined model updates.

The application discloses a digital asset circulation method and system based on multi-end linkage relationship chain right confirmation, relates to the technical field of digital asset processing, and comprises the following steps: acquiring a multi-end concurrent submission right confirmation edge operation request set, determining a relationship chain graph atlas local topology change record according to a source node identifier and a target node identifier; cross-comparing an added right confirmation edge operation request and a deleted right confirmation edge operation request in the local topology change record to determine mutual exclusion operation pairs for the same right confirmation edge; and determining a right confirmation edge final retention state based on a local timestamp attached to the mutual exclusion operation pairs and updating the right confirmation edge final retention state to a combined relationship chain graph atlas.

The application belongs to the technical field of electric power automation, and discloses a power grid power flow calculation method based on a graph neural network and related devices; the method comprises the following steps: obtaining power grid data to generate a power flow data set; constructing node feature vectors and edge feature vectors based on the power flow data set; splicing the node feature vectors and the position encoding of the nodes in the power grid to obtain spliced features; inputting the spliced features and the edge feature vectors into a pre-trained graph neural network model to obtain a power grid power flow calculation result based on the graph neural network; compared with the traditional Newton-Raphson method and the graph neural network method considering local topological relations, the application can significantly improve the calculation speed while maintaining the accuracy and reliability of the result; the application also has beneficial effects such as strong adaptability and wide application scenarios, can adapt to power grids of different scales and topological structures, has good generalization ability, and can be applied to various application scenarios such as real-time power flow calculation and future state power flow prediction.

The invention discloses a leakage point location positioning method and device for a water supply network and a medium, belongs to the technical field of operation and maintenance of the water supply network and leakage control, and aims to solve the problems that in existing leakage point positioning, single data is prone to being depended, devices and algorithms are complex, high-precision space linkage is lacked, and data adaptability and refinement are insufficient. And accurate positioning of the water leakage point of the water supply network is difficult to realize. The method comprises the following steps: carrying out multi-source data acquisition processing on the running state of the water supply network to obtain core data of the water supply network; performing feature vector standardization processing on the preprocessed corrected water supply network core data to obtain effective water supply network data; leakage point area preliminary screening under the related leakage point risk value is carried out on the effective water supply pipe network data, and a highly-suspected leakage point area based on a water supply pipe network road section is obtained; and carrying out leakage point space positioning calculation under local topology on the highly-suspected leakage point region to obtain a final leakage point region position in the water supply network grid.

The invention discloses a quadruped robot three-dimensional map adaptive compression method under a complex terrain, and the method comprises the steps: extracting Z-axis height information of each key frame in a three-dimensional point cloud map, and obtaining an original height sequence; filtering and smoothing the original height sequence; dividing the key frame into a plurality of compression units based on the smooth height sequence; for each compression unit, identifying an initial ground point through an RANSAC plane fitting method, removing non-ground points in combination with normal vector constraint and a height threshold, and determining a reference ground height according to optimized ground points; setting a compression height range based on the reference ground height, and taking the three-dimensional point cloud in the compression height range as a to-be-projected point set; performing orthographic projection on the to-be-projected point set to generate a two-dimensional local sub-graph; and splicing the two-dimensional local sub-graphs to generate a global two-dimensional occupied grid map. According to the method, on the premise that the local topological relation and navigation availability are kept, the complex three-dimensional space is flattened into the two-dimensional map in a segmented mode.

The invention discloses a reconfigurable battery topological structure and a battery energy storage system balance control method thereof, and relates to the field of battery energy storage, in a hardware level, each battery unit is configured with a single-pole double-throw switch and a single-pole single-throw switch connected with a parallel bus through a current-limiting resistor. In the software level, a battery system is abstracted into a graph structure, and a graph neural network (GNN) is used for extracting local topological features and fusing global information. Besides, modeling is carried out on action probability distribution in combination with multi-agent reinforcement learning, a security constraint verification and instruction correction mechanism is introduced, it is ensured that the finally generated switch control signal can achieve an almost optimal equalization effect in a huge state space and strictly accords with a hardware security criterion, and the method is suitable for the large-scale industrial control. Therefore, the available capacity, the reliability and the service life of the energy storage system are comprehensively improved.