122 results about "Multiple-scale analysis" patented technology

The invention discloses an aero-engine residual life prediction method based on multi-modal deep learning, and relates to the field of aero-engine prediction and health management. The method comprises the following steps: acquiring and preprocessing multi-sensor time sequence data; constructing a degradation sensitive feature set through multi-scale analysis of a time domain, a frequency domain and a time-frequency domain; constructing a multi-modal deep learning model comprising an original data processing module and a multi-scale feature processing module, and introducing an attention mechanism and an uncertainty quantization module into the model; the model is subjected to lightweight processing to support embedded deployment. According to the method, multi-scale features and multi-modal deep learning are fused, the uncertainty quantification capability is achieved, high-precision and interpretable residual life prediction with uncertainty quantification is achieved, and reliable support is provided for engine maintenance decision making.

The invention discloses an intelligent geophysical exploration geological exploration analysis system, and relates to the field of geological exploration, and the system comprises a collection and preprocessing module which is used for carrying out the adaptive collection of multiple types of geophysical field signals of a target exploration region, and synchronously carrying out the preliminary noise reduction and quality optimization of original signals through signal purification processing; the fusion module is used for receiving the preprocessed multi-type geophysical field signals and carrying out normalization and feature depth correlation fusion on coordinates of data of different dimensions through spatial-temporal feature correlation processing; the method can dynamically adapt to environmental noise optimization signal acquisition, extracts stratum recessive characteristics through multi-dimensional data fusion and intelligent mining, constructs an accurate three-dimensional geologic model, positions stratum attribute deviation and grades through multi-scale analysis, presents exploration results in combination with three-dimensional visualization, and generates target region optimization suggestions.

The invention discloses a composite material performance prediction method based on multi-field coupling multi-scale analysis and a knowledge graph, which integrates microstructure modeling, graph neural network representation learning, cross-scale parameter coupling modeling and robust optimization analysis. The method is suitable for prediction and design of key mechanical properties such as modulus, strength and toughness of a thermosetting / thermoplastic composite material under different working conditions. The method comprises the following steps: firstly, constructing a grain-level tissue knowledge graph based on an electron backscatter diffraction image, secondly, constructing a multi-scale input system comprising a microscopic variable (such as a fiber volume fraction), a mesoscopic variable (such as a layer thickness sequence) and a macroscopic variable (such as a load condition), and finally, performing robust optimization by utilizing a multi-objective evolutionary algorithm to obtain a multi-scale input system. And outputting a material performance prediction result and a knowledge graph associated with the ji structure-process-performance. According to the method, the accuracy and interpretability of performance prediction of the composite material can be remarkably improved, and data-knowledge dual-drive support is provided for design optimization of the high-performance composite material.

The invention relates to the technical field of network security, in particular to a security monitoring method and system for network traffic. The method comprises the following steps: capturing a network flow data flow in real time, and extracting a network entity and a direct communication relationship to construct a basic communication graph; identifying and quantifying a high-order interaction mode between network entities, and taking the high-order interaction mode as an implicit feature enhanced basic communication graph to generate an enhanced security graph; processing the enhanced security map by using a multi-scale time sequence diagram neural network, and capturing a short-term burst mode and a long-term evolution mode at the same time; a dynamic anomaly score is calculated based on the network entity historical behavior baseline and the current network situation, and a security alert is generated when an adaptive threshold is exceeded. The system correspondingly comprises a flow capture module, a feature extraction module, a high-order mode analysis module, a security map construction module, a multi-scale analysis module, a dynamic risk assessment module and an intelligent alarm module, and comprehensive and accurate network threat detection is realized.

The invention discloses an abnormal transaction dynamic detection method and system fusing multi-scale analysis and information entropy, and the method comprises the steps: S01, constructing a dynamic heterogeneous graph which comprises account nodes, equipment nodes and IP nodes; s02, extracting node features of each node, extracting a spectrum feature matrix of the dynamic heterogeneous graph, and forming a node feature vector of each node; step S03, clustering the dynamic heterogeneous graph to screen out candidate detection sub-graphs, inputting node feature vectors of nodes in the candidate detection sub-graphs into a pre-trained graph classification model, and identifying an abnormal account node set; and S04, calculating a corresponding dynamic risk score in real time according to the memory state vector of each abnormal account node in the time sequence diagram network and the diagram information entropy change rate so as to evaluate a real-time risk and identify a core abnormal account. According to the method, the group type abnormal transaction account can be quickly and accurately identified, and the core abnormal account can be positioned.

The invention relates to the technical field of battery manufacturing, and discloses an environment-friendly manufacturing method of a high-capacity battery. The method comprises the following steps: firstly, collecting relevant data of sustainable material attributes, green process monitoring and environment footprints, and extracting an initial feature time sequence from the relevant data; a fusion feature set of material-process-environment interaction is extracted, a multi-scale model structure is constructed in combination with data acquisition precision of different making partitions, and multi-resolution feature mapping is generated; performing trend analysis on the time evolution data to generate a battery performance degradation spectrogram; stress response distribution of the core production stage is determined through simulation calculation, and a manufacturing state response data set is obtained; and identifying the number of performance degradation cycles, calculating a residual capacity value by fusing a material degradation model, and evaluating the total life of the battery produced in an environment-friendly manner after integration. According to the method, environmental protection and green consideration of the whole production process is realized through multi-dimensional data fusion and multi-scale analysis.

The invention discloses a lightweight real-time mapping system based on cloud edge collaboration and a VLA-HDMap high-frequency interaction method, and aims to solve the problems of slow updating, high cost and low safety of HDMap. The model has the capabilities of real-time environment understanding, semantic analysis and end-to-end decision instruction output. The vehicle end can download the cloud HMap in real time, and also participates in crowdsourcing data acquisition and uploading. According to the local lightweight mapping method provided by the invention, a real-time circulation mechanism of'acquisition-application-correction 'is used, so that the HDMap is highly consistent with the actual environment. A cloud edge collaborative hierarchical parallel incremental update global HDMap is also provided, the essence is a time-space intersection multi-scale analysis model of minute-level traffic flow prediction and day / week-level map differentiation detection, a high-risk region of traffic flow peak + map change can be efficiently identified, and a convex hull algorithm is used to obtain a region, which needs to be locally updated, of the whole HDMap. According to the invention, a new-generation intelligent traffic system can be constructed, and the safety application of group intelligence, symbiotic economy and large-scale automatic driving can be promoted.

The invention relates to the technical field of water body pollution monitoring, in particular to a multispectral image fused water body pollutant accurate recognition system, which comprises a data acquisition module, a cloud processing module, a boundary processing module, a pollution recognition module, a diffusion prediction module and a visualization module, the system constructs sub-pixel representation of a water pollutant boundary by using a differential geometry manifold theory, and realizes high-precision pollutant boundary description through a multi-scale analysis and curvature flow optimization technology; enhancing pollutant characteristic expression by adopting multispectral image fusion and an optimal wave band selection technology; using a support vector machine model to accurately identify various pollutant types such as oil films, oil spots, algae blooms and the like; based on a boundary fine description result and a pollutant type identification result, the diffusion trend of pollutants is accurately predicted in combination with historical flow, wind direction and wind speed data, and the system improves the water pollutant boundary identification precision and enhances the identification capability of complex boundary forms and low-contrast regions.

The invention discloses a fire behavior detection method based on multi-modal large model multi-scale analysis and depth reasoning, and relates to the technical field of artificial intelligence safety monitoring. The method comprises the following steps: acquiring visible light video data, infrared image data and audio data in the same fire scene; respectively preprocessing the data to obtain processed data; inputting the processed data into a corresponding modal encoder in a pre-trained ImageBind multi-modal model for processing, and respectively outputting a feature corresponding to each modal; a gating fusion module is adopted to perform fusion calculation on the features corresponding to the modals, and unified scene representation is output; and carrying out multi-dimensional depth reasoning by adopting a depth reasoning module to obtain a comprehensive result of the multi-dimensional reasoning, and outputting a fire confidence coefficient score and an interpretability basis based on the comprehensive result of the multi-dimensional reasoning. According to the invention, the accuracy of fire detection can be improved.

The invention discloses a distribution box operation state monitoring method and device, and relates to the field of intelligent operation and maintenance of a power system. The method comprises the following steps: firstly, processing collected environment, load and vibration acceleration data of a distribution box to obtain standardized multi-dimensional time series data; monitoring characteristic parameters are extracted from the data, and short-term and long-term multi-scale characteristic vectors are obtained through sliding window statistics; and finally, inputting the data into an attention mechanism feature fusion model to obtain a risk anomaly score, setting a dynamic threshold value by combining statistical distribution of historical normal operation scores of the distribution box, and judging a fault trend type according to the dynamic threshold value. The scheme has the advantages that the limitation of a single data source is avoided through multi-source data fusion, and the anomaly recognition accuracy is improved; single window information and long-term trend can be captured through multi-scale analysis; the dynamic threshold value adapts to different equipment and environments and is more accurate than a fixed threshold value; and hierarchical evaluation can provide clear maintenance suggestions and assist operation and maintenance decisions.

The invention discloses a multi-source state trend integrated data fusion method for a weak support large-scale new energy delivery system. According to the invention, a data source identification and hierarchical access mechanism, a multi-source synchronization and calibration method based on a unified time reference, a topology-oriented space mapping method, a trend feature extraction method based on a sliding window and multi-scale analysis, and a state-trend integrated weak support risk characterization index system are constructed; unified modeling and online sensing of key quantities such as voltage supporting capacity, power flow margin, standby level, renewable output fluctuation and the like are realized. Compared with an existing monitoring mode depending on a single system, a single time scale or a single index, the method has the advantages that on the basis of ensuring data quality and time-space consistency, a unified data base and trend sensing capability can be provided for safety evaluation, cooperative regulation and market linkage control of the weak-support large-scale new energy delivery system; and the method has relatively high engineering applicability and popularization value.

The invention discloses a rock particle interface deformation identification method based on an image identification algorithm, and belongs to the field of rock particle interface deformation identification, and the method comprises the steps: obtaining a particle distribution range through a high-resolution image collection and multi-scale scanning technology, and when the scale span is detected to exceed a preset threshold value, determining the particle distribution range; and separating particle regions with different sizes by adopting a layered segmentation method. The texture mode of each layer is analyzed by using a convolutional neural network, a deformation feature vector is extracted, and algorithm granularity parameters are adjusted through iterative optimization so as to match the actual particle size. And when the matching degree is insufficient, a feedback loop mechanism is introduced to reprocess the image data and a noise filtering link is fused, so that the deformation identification precision is remarkably improved. Finally, an integrated calculation method is adopted to fuse multi-scale analysis results, accurate judgment of the deformation state of the whole particle interface is achieved, reliable comprehensive indexes are provided for rock stability evaluation, and the technical problem that a traditional method is insufficient in precision in cross-scale particle deformation recognition is solved.

The invention provides an intelligent underground space ground surface settlement remote sensing identification and evaluation method, and aims to solve the problems of low monitoring efficiency, insufficient precision and lack of comprehensive risk evaluation in the prior art. Multi-scale analysis space-time filtering is adopted to suppress the atmospheric phase; then constructing an entropy weight-principal component analysis fusion model to realize multi-source feature collaborative fusion, and applying a convolutional neural network (MSA-CNN) introduced with a multi-scale space attention module to intelligently identify a settlement area and optimize a boundary; and finally, quantifying a risk level based on an improved entropy weight-fuzzy comprehensive evaluation model, carrying out trend prediction and early warning by adopting an attention-enhanced LSTM model, and integrating a result to a WebGIS platform. According to the invention, full-process automatic monitoring is realized, the false alarm rate is reduced by more than 20%, and scientific decision support is provided for safety management of underground engineering.

The invention discloses a multi-scale analysis-based gas-water seepage field simulation method, which specifically comprises the following steps of: based on a digital rock core reconstructed by a CT (Computed Tomography) scanning image of a typical rock sample of a target reservoir, adopting a microcosmic pore network flow simulation technology; the change rule of the stratum seepage capacity in different stratum pressure gradients and water drive processes, the use degree of a pore network and the water drive process in a pore are calculated, and then a correlation chart with the rock permeability and the pressure gradients is established and used for rapid measurement and calculation of the actual reserve use degree and the recovery degree of the water-invaded sandstone gas reservoir. The method starts from a rock microcosmic pore channel, the scale is amplified to a macroscopic oil reservoir during application, the method has the cross-scale characteristic, and the method is used for rapidly measuring, calculating and evaluating the reserve utilization degree and the recovery degree of the actual water-invaded sandstone gas reservoir and is beneficial to rapid development of the gas reservoir engineering scheme design in a mine field.

The invention belongs to the field of water environment monitoring and water quality prediction, and particularly discloses a causal discovery and deep learning coupled water quality prediction method and system, and the method comprises the steps: carrying out the causal screening of dynamic covariables through employing a causal discovery algorithm based on a neural network, and recognizing causal dynamic covariables; and inputting the causal dynamic covariable, the multi-scale water quality index characteristics and the static covariable data into a trained probability time sequence prediction model, and outputting a probabilistic prediction result of the water quality index concentration of the target water area in a plurality of time steps in the future. According to the method, pseudo-correlation variables are eliminated, the non-stationarity is processed through multi-scale analysis, purer and richer multi-scale information is provided for the probability time sequence prediction model, and the accuracy of point prediction is remarkably improved.

The invention discloses a method for evaluating the residual strength of a multi-scale fiber-reinforced flexible pipe under a stretch-bending load, and relates to the technical field of structural strength evaluation of ocean engineering composite materials, and the method comprises the following steps: extracting fiber distribution characteristics by adopting digital image processing based on a scanning electron microscope image; a Monte Carlo method is combined to construct a two-dimensional RVE geometric model of the fiber reinforced structure containing the damp-heat degradation, and macroscopic elastic parameters are predicted; a VUMAT user material subprogram containing dynamic damage is written based on Fortran, damage starting and evolution failure criteria of a fiber reinforced structure are embedded, and the nonlinear coupling evolution process of multiple damage modes such as fiber fracture, matrix cracking and interface debonding is simulated. According to the method, a material-damage-response multi-scale analysis framework is constructed by coupling microscopic material components, a microscopic damage model and macroscopic pipeline structure analysis, the problem that a traditional method lacks multi-scale damage coupling analysis under a complex load is solved, and the residual strength evaluation precision of the pipeline under the action of a stretching-bending combined load is improved.

The invention discloses a composite material multi-scale image finite element analysis method considering manufacturing characteristics, and belongs to the technical field of special-shaped composite material structure multi-scale analysis. CT scanning is carried out on the special-shaped composite material structure, an internal image of the special-shaped composite material structure is obtained, statistical parameters of manufacturing characteristics are extracted through an image processing technology, and mesoscopic representative volume units are classified according to the statistical parameters; secondly, a self-consistent clustering analysis method is adopted for model unified order reduction, and an efficient solving database is constructed; mapping corresponding cell element categories according to defect characteristics of each region of the component, realizing scale solution through a concurrent multi-scale framework and a homogenization method, and predicting mechanical response and damage evolution of the structure; and finally, revealing a physical mechanism that macroscopic damage expands along a low-performance weak band based on spatial distribution of cell categories. The method has high calculation efficiency and prediction authenticity, and an effective tool is provided for performance evaluation and process optimization of the defect-containing composite material component.

The invention discloses a method and a system for automatically analyzing OTDR (Optical Time Domain Reflectometry) test curves of a hollow-core optical fiber. The method comprises the following steps of: acquiring one OTDR test curve from each of two ends of the hollow-core optical fiber; respectively carrying out centralization processing on the two OTDR test curves, aligning the two OTDR test curves based on the position of the hollow-core optical fiber, calculating a difference value of backscattering power of the two OTDR test curves, and then calculating an average value to obtain a correction curve; performing filtering and noise reduction processing on the corrected curve to obtain a curve after noise reduction; extracting event information from the denoised curve by adopting a multi-scale analysis and dynamic threshold fusion algorithm; and exporting an event detection result according to the event information. According to the invention, automatic processing and analysis of the hollow-core optical fiber OTDR curve are realized while bump interference is avoided.

The application discloses a kind of hollow optical fiber OTDR test curve automatic analysis method and system, and the OTDR test curve of 1 piece is collected from the two ends of hollow optical fiber;2 OTDR test curves are respectively centered, and the backscattering power of 2 OTDR test curves is difference after being aligned based on hollow optical fiber position, and then average is obtained, to obtain correction curve;Filtering noise reduction processing is carried out to correction curve, to obtain the curve after noise reduction;Multi-scale analysis and dynamic threshold fusion algorithm are used to extract event information from the curve after noise reduction;According to event information, event detection result is derived.The application realizes the automatic processing and analysis of hollow optical fiber OTDR curve while avoiding bulge interference.

The invention relates to the technical field of safety helmet detection, in particular to a safety helmet detection system based on multi-scale image generation and analysis, an image layering module obtains multi-scale images of a scene to be detected and a standard safety helmet, performs feature decomposition and extracts multi-scale feature components; the difference modeling module analyzes the change coefficient and the average level of the key point position difference, and generates a background interference factor; the consistency evaluation module evaluates the feature consistency degree according to the feature intensity difference fluctuation condition; the interference judgment module fuses the interference factors and the consistency degree, constructs a superimposed interference feature value, screens effective feature vectors and fuses the effective feature vectors to obtain target features after background interference elimination; and the target detection module judges the existence and category of the safety helmet by using a classifier based on the optimized features. The system effectively suppresses complex background interference through a multi-scale analysis and feature quality evaluation mechanism, improves the detection precision and robustness in a small-target and multi-scale scene, and is suitable for industrial safety monitoring application.

The invention discloses a precipitation-driven river pollution source analysis method based on multi-scale analysis and application. The analysis method comprises the following steps: acquiring precipitation data, river water quality data and hydrological data of a to-be-researched region for many years; determining a pollution type according to the river water quality data sudden change point; based on the pollution type, positioning pollution time and a pollution area by adopting a space-time heterogeneity analysis method; aiming at the pollution time and the pollution area, establishing a relation between rainfall and river water quality pollution. According to the method, the analysis limitation of a traditional single spatial-temporal scale is broken through, and point-line-plane multi-dimensional response analysis is achieved; through coupling modeling of water quality mutation recognition and rainfall driving analysis, the superimposed influence of human activities and meteorological hydrology on river water quality is systematically revealed, and comprehensive analysis of the relationship between rainfall and river water quality pollution is realized in combination with year, season and day multi-scale collaborative analysis.

The invention discloses a machine vision-assisted optic nerve injury repair state detection method, and relates to the technical field of medical informatics, and the method comprises the steps: through interaction with a front-end equipment group, determining a dual-mode image, triggering a state detector, executing multi-scale image reconstruction and A2 polarization reasoning under dual-path image feature extraction, and determining an A2 polarization probability heat map; the method comprises the following steps: evaluating a repair state, performing directional detection guidance, performing multiple rounds of detection analysis based on first-order global detection and second-order directional detection in a repair period until the repair period is finished, and generating a periodic state evolution graph, so as to solve the problems that in the prior art, the detection is one-sided, and repair associated information of different levels cannot be effectively applied; the technical problem that the detection intelligence and the accuracy of repair state evaluation are limited due to multi-modal imaging information in the prior art is solved, and the intelligent degree and the accuracy of detection can be effectively improved by adopting a systematic detection mode which can fuse multi-modal imaging information and has multi-scale analysis and intelligent reasoning capabilities.

The invention relates to a vegetation leaf dry matter content remote sensing high-precision inversion method. The method comprises the following steps: S1, constructing a vegetation leaf sample data set; s11, constructing a blade actual measurement data set; s12, generating an analog data set and / or an analog data set added with noise; s13, dividing the actual measurement data set into an actual measurement training set and an actual measurement verification set; s2, dry matter weak information features are extracted through continuous wavelet transform; s21, carrying out multi-scale analysis calculation on dry matter weak information by using a continuous wavelet transform method; s22, performing wavelet basis function transformation on the original reflection spectrum of each leaf sample to obtain wavelet coefficient characteristics; s23, carrying out correlation analysis calculation on the wavelet coefficient characteristics and the LMA; s24, a threshold value is set, and wavelet coefficient characteristics with sensitivity to LMA spectrum weak information are screened out; and S3, constructing an LMA inversion model based on wavelet coefficient coupling machine learning. The method is high in inversion precision and strong in noise robustness.

The application discloses a kind of electric energy spectrum analysis methods based on combination sparse conversion.The method steps are: real-time monitoring electric power signal in power system and pre-processing, different scale frequency domain wavelet coefficients are obtained using fast Fourier transform, then frequency domain wavelet coefficients are optimized by time-frequency coupling method, then time-domain signal is reconstructed, and it is converted back to frequency domain using integer discrete Fourier transform, then frequency spectrum adaptive adjustment is carried out to frequency domain signal, and the frequency distribution of optimization is carried out, finally, frequency spectrum analysis is carried out based on adjusted frequency domain signal.The application improves the frequency spectrum resolution of electric energy signal by combining fast Fourier transform and integer discrete Fourier transform, and using multi-scale analysis and time-frequency coupling optimization technology, realizes high-precision frequency spectrum analysis and interference suppression, has the advantages of anti-dynamic current interference, adaptive enhancement key frequency band and the advantages such as improvement of signal-to-noise ratio.

The invention relates to the technical field of e-commerce logistics management, and discloses an e-commerce logistics tracking system based on a block chain. The system obtains an order identifier through a distributed query module, retrieves logistics records in a block chain historical database, and assembles a historical logistics record set. And the event distinguishing module carries out event type distinguishing processing on the record set to generate a heterogeneous logistics event set. The feature extraction module executes feature extraction for each event set to obtain a logistics feature core value set and an associated time window set. And the multi-scale analysis module performs multi-level analysis on the logistics data sequence through the block chain smart contract by using the associated time window set as an analysis reference, and generates a plurality of logistics feature groups. And the feature fusion module performs inter-scale fusion processing on the feature group, and outputs a final fused logistics feature group as a logistics tracking result. According to the system, the transparency and credibility of e-commerce logistics tracking are improved, and multi-dimensional and multi-scale analysis of the whole logistics process is realized.

The invention discloses a correlation analysis method for railway slope deformation characteristics under a rainfall condition, and relates to the technical field of slope deformation, and the method comprises the steps: S1, carrying out the observation investigation of a target region and a target slope through an unmanned plane and on-site investigation, and analyzing the morphological characteristics and position characteristics of the slope; s2, analyzing the deformation rate and space-time deformation characteristics of the target area by using an InSAR technology, and then denoising original InSAR deformation data by using Gaussian filtering GF and wavelet transform WT; carrying out interpolation processing on the time sequence deformation features by adopting a plurality of data continuity algorithms; and S3, comprehensive analysis of deformation-rainfall correlation: firstly carrying out correlation analysis on rainfall and deformation sequences of each month, and then calculating annual correlation between deformation and rainfall characteristics. The invention solves the problem that technologies for multi-scale analysis of railway slope deformation and quantitative analysis of correlation between slope deformation and rainfall are deficient, and provides a new method for railway slope rainfall-deformation feature and correlation analysis.

The application relates to the technical field of network analysis, and discloses a network community discovery system and method through matrix analysis, which comprises the following modules: a network module, which converts a static network topology into an information propagation model, and establishes a node state time sequence dynamic model by defining a node information processing rule and a multi-round propagation mechanism; a phase space reconstruction module, which maps high-dimensional time sequence data to low-dimensional phase space through a nonlinear dimension reduction method to form node trajectory distribution data; a community feature module, which analyzes the convergence, oscillation mode and attractor feature of the node trajectory, and generates community structure feature data through trajectory similarity; and a community division module, which identifies a community boundary through density clustering, and constructs a hierarchical community organization through multi-scale analysis. The application can deeply mine the internal community structure of a network from the perspective of dynamic information propagation, and overcomes the limitation that traditional methods only consider static topology.

The application discloses a point cloud local curvature feature calculation method and system for structural surface identification, and particularly relates to the technical field of three-dimensional point cloud data processing and feature extraction, and is used for solving the problem that the existing method leads to local curvature feature distortion due to direct calculation based on original point cloud containing noise, and then affects the subsequent structural surface identification precision; a plurality of different neighborhood radiuses are calculated respectively for the query point in the point cloud, and a plurality of sets of eigenvalues and eigenvectors are obtained by performing feature decomposition on the covariance matrix; an optimal set is selected based on the stability and consistency of the feature structure, and then the curvature tensor of the point is calculated; the candidate local curvature eigenvalue is derived from the curvature tensor and verified and output according to the geometric flatness prior knowledge of the structural surface; through the multi-scale analysis and verification mechanism, the interference of noise and local fluctuation is effectively inhibited, and more reliable and more robust local curvature features can be calculated.

The invention relates to the technical field of geological exploration, in particular to an on-site exploration stratum logging method and system. The method comprises the following steps: identifying a layered boundary in a rock core image by using a deep learning image segmentation model; extracting a stratum interface in the logging curve through wavelet transform and multi-scale analysis; based on a multi-source feature fusion strategy of an attention mechanism and an auto-encoder, the stratigraphic boundary, the stratigraphic interface and the seismic facies feature vectors are combined to analyze comprehensive features of the stratigraphic boundary; according to a stratum attribute graph construction strategy of the graph neural network, associating a stratum reflection layer and the lithology classification feature vector in the radar image to obtain stratum attribute graph features; and constructing a comprehensive decision-making model of a random forest classifier and a rule engine, and generating a stratum catalog report by combining the landform boundary, the stratum boundary comprehensive features and the stratum attribute graph features. According to the method, the accuracy, objectivity and efficiency of the catalog result are improved by cooperatively processing the multi-source data.

The invention discloses a data acquisition pre-filtering processing system based on three-dimensional laser scanning. The data acquisition pre-filtering processing system comprises a multi-modal data acquisition module which is used for synchronously acquiring geometric characteristics, multispectral intensity data, polarization parameters and positioning and attitude determination information; the multi-modal vegetation penetration preprocessing module is used for outputting pure terrain point cloud and vegetation masks; the terrain self-adaptive dynamic parameter optimization module is used for acquiring pure terrain point clouds as source point clouds and target point clouds, searching candidate corresponding points in the target point clouds for each point in the source point clouds, performing multiple filtering on the candidate corresponding point pairs, performing coordinate transformation on the source point clouds by applying the optimal rigid body transformation matrix, and performing dynamic parameter optimization on the target point clouds; calculating a registration error between the transformed point cloud and the target point cloud; the micro-variation feature enhancement multi-scale analysis module is used for receiving the dynamically optimized parameters and carrying out multi-scale pyramid decomposition and feature saliency detection on the pure terrain point cloud; and the quality evaluation and output module is used for carrying out quality evaluation on the processed point cloud.