50 results about "Geometric matching" patented technology

The invention discloses an industrial automobile crane positioning and mapping method and device based on multi-sensing fusion, and relates to the field of machine sensing, and the method comprises the following steps: initializing a pose and offset of an inertial measurement unit by using an input laser point cloud and a pre-integration result of the inertial measurement unit; correcting laser point cloud motion distortion by using an inertial measurement unit, extracting a dynamic object point cloud, and synchronously extracting laser features and visual features with depth; calculating a key frame pose in real time by minimizing a visual re-projection error and an inertial measurement unit pre-integration error, and performing frame-to-map matching by taking the key frame pose as an initial value to further optimize the pose; a visual word bag model is utilized to quickly retrieve a closed-loop candidate frame, after laser passes geometric matching to verify a closed loop, a laser radar factor, a visual factor, an inertial measurement unit factor and a closed-loop factor are fused, a global factor graph is constructed and optimized, and a pose and point cloud map is updated. According to the method, the environment sensing capability of the industrial automobile crane can be improved, and the sensing accuracy and robustness are improved.

The invention discloses a physical object three-dimensional positioning method and system based on video data, and relates to the technical field of three-dimensional positioning. The method is used for solving the problems of low object space positioning precision, unstable pose estimation and poor time sequence continuity in a video scene. Firstly, an input video stream is analyzed, an object segmentation mask is generated, feature points are extracted, camera motion parameters are calculated through inter-frame matching, and a scene sparse three-dimensional point cloud is reconstructed; a candidate three-dimensional bounding box is generated according to the segmentation mask and the point cloud, an optimal bounding box is selected by combining geometric matching degree and feature similarity evaluation, and a preliminary three-dimensional positioning result of the object is obtained; a positioning result and historical frame motion data are fused, a space-time constraint optimization model is constructed, and the six-degree-of-freedom pose of the object is solved; and finally, neural radiation field representation is established based on the pose, the pose and neural parameters are optimized through combination of micro rendering and back propagation, continuous and accurate updating of the three-dimensional position of the object is realized, and the positioning stability and robustness in a dynamic scene are improved.

The invention provides a Bayesian fusion-based 4D millimeter wave radar point cloud enhancement and target detection method, and the method comprises the steps: 1, projecting a 4D millimeter wave radar point cloud to an image plane, carrying out the matching of an image instance segmentation result, and obtaining an image semantic category and image segmentation confidence; 2, constructing a class conditional probability density model based on the physical characteristics of the radar points; based on a Bayesian probability framework, carrying out weighted fusion on the image segmentation confidence and the class probability density to obtain fusion confidence; 3, taking the fusion confidence and semantic category codes as augmented features, and performing feature splicing on the augmented features and the 4D millimeter-wave radar point cloud to generate a semantic enhanced point cloud; and sampling the semantic enhanced point cloud by adopting a sampling method fusing confidence guidance, and inputting the sampled semantic enhanced point cloud into a detection network for training and reasoning. According to the method, the technical problems of unidirectional dependence, shallow geometric matching and insufficient utilization of radar physical characteristics in the existing image and radar point cloud fusion technology are solved.

The invention belongs to the technical field of building construction and digital twinning, and relates to an intelligent matching method and system based on artificial intelligence. According to the method, construction environment data are collected in real time, risk marks are identified, the deformation amount of the component is calculated in combination with a time sequence prediction model, and a BIM model is corrected to generate a virtual component for environment compensation; carrying out geometric matching analysis by adopting a three-dimensional point cloud registration technology, and judging a hole alignment error and an interface gap; dynamically adjusting the installation sequence based on the dynamic decision tree model by integrating the geometric mismatch degree, the environmental risk and the construction logic; and carrying out incremental updating on the model according to the actual installation deviation. According to the method, the technical problem of installation mismatch caused by component deformation due to environmental factors such as temperature and humidity in the construction process is effectively solved, dynamic adaptation and precision control of the construction process are achieved, the matching success rate and construction efficiency of component assembly are improved, and meanwhile the adaptability and reliability of the system are optimized through continuous learning.

The invention discloses a small sample non-uniform multi-mode clutter modeling and partition suppression method, which is applied to the technical field of radars, performs multi-mode clutter modeling aiming at a complex environment, partitions the non-uniform clutter environment based on the difference of non-uniform clutter covariance matrix geometric matching, and improves the robustness of the multi-mode clutter modeling. Then, the problem that the performance of a traditional clutter suppression method is reduced under the small sample condition is solved through a dimensionality reduction self-adaptive filtering method in different areas; according to the method, firstly, multi-mode clutter echoes are generated for a complex environment, different-mode clutter region boundaries are judged through the difference of different clutter covariance matrix geometric matching, clutters are partitioned, and then the degree of freedom is reduced through data dimension reduction in different regions, and filter weight vectors are designed to achieve self-adaptive filtering; according to the method, effective partitioning and suppression of multimode clutters can be realized under the conditions that the number of samples is less than the degree of freedom of a radar system and the clutter environment is not uniform.

The invention discloses an intelligent blanking method for manufacturing a novel composite material for a retired wind power blade, and relates to the technical field of signal processing. The method comprises the following steps: acquiring geometric data of a blade and degradation characteristics of an internal material through a three-dimensional laser scanning and nondestructive testing technology, and generating a high-precision material non-uniformity distribution map; calculating a fiber feature matching degree and a stress buffer margin based on the atlas in combination with performance requirements of the composite material, and generating a reinforced phase distribution scheme; a geometric matching degree and a cutting angle compensation value are calculated according to the scheme, and a cutting track with optimized stress distribution is planned; the cutting equipment is driven to execute the track, the acoustic emission vibration spectrum is monitored in real time, and a control instruction containing a self-adaptive speed reduction mechanism is generated; and performing closed-loop feedback on the quality data based on the blanking unit to the atlas construction step. According to the method, the material utilization rate maximization, fiber continuous protection and defect active avoidance are realized, and the performance stability and the production efficiency of the composite material are remarkably improved.

The invention discloses an unmanned aerial vehicle landslide image clustering method and device for emergency monitoring, and belongs to the technical field of landslide monitoring. The method comprises the following steps: acquiring an original image shot by a camera mounted on an unmanned aerial vehicle; extracting multi-modal features of the landslide instance from the original image, wherein the multi-modal features comprise a spatial position feature, a global semantic embedding feature and a local geometric matching feature; fusing the multi-modal features by using a neural network model based on an attention mechanism to generate an instance feature vector; on the basis of the spatial position features, the local geometric matching features and the instance feature vectors, through an iteration process including a feedback mechanism, a graph structure representing the incidence relation between the landslide instances is constructed; and connected components in the graph structure are extracted, and each connected component is determined as a landslide instance cluster. According to the invention, different image instances belonging to the same landslide can be accurately associated, and the error rate of cross-frame target matching under a complex terrain is reduced.

The invention discloses an apple three-dimensional reconstruction method and device under a limited visual angle, a storage medium and equipment, and the method comprises the steps: obtaining a multi-visual-angle image which corresponds to a to-be-reconstructed apple and only covers a part of a visible surface, and carrying out the three-dimensional reconstruction of a visible surface region of the apple based on the multi-visual-angle image through a three-dimensional reconstruction model, determining a target three-dimensional point cloud of the apple; extracting a local theoretical point cloud corresponding to the target three-dimensional point cloud from the Apple theory geometric model, and performing space alignment on the local theoretical point cloud and the target three-dimensional point cloud; according to a spatial alignment result, constructing a geometric matching error function between the local theoretical point cloud and the target three-dimensional point cloud, and based on the geometric matching error function, optimizing the Apple theoretical geometric model by taking reduction of geometric matching errors as an optimization target so as to obtain an optimized Apple theoretical geometric model; and optimizing the three-dimensional reconstruction model by using the optimized apple theoretical geometric model, and performing three-dimensional reconstruction covering the complete surface of the apple.

The invention provides a laser scanning historic building structure data acquisition method and system, relates to the technical field of laser scanning, and can calculate the scanning range of a scanner in real time by obtaining a preset key monitoring area in a historic building and the real-time spatial position and attitude information of the laser scanner. On the basis, a three-dimensional bounding box intersection algorithm is adopted to carry out real-time space geometric matching calculation on the scanning range and the key monitoring area, so that whether an overlapping relation exists or not is judged. And intelligently marking the collected point cloud data as high-value data or common-value data according to the overlapping relationship. According to the method, the high-value data is written into the high-priority cache queue, the common-value data is written into the common-priority cache queue, and differentiated cache management and unloading strategies are executed on the data in the two queues, so that the dynamic monitoring task of the historic building can be effectively carried out; and the problem that the model loses the analysis value due to data missing or timestamp disorder is avoided.

This invention discloses a cross-view positioning method based on semantic coarse localization and BEV fine registration, aiming to solve the problem of large-scale, high-precision global positioning under conditions of no initial position prior. The invention employs a two-stage processing architecture: First, hierarchical semantic information is extracted from multi-view images from the vehicle using a visual language model. Combined with heading constraints and a backtracking retrieval strategy, candidate regions are quickly screened in a standard map to achieve semantic coarse localization. Then, within the candidate regions, the vehicle-mounted BEV perception results are finely geometrically registered with the map template, and a dynamic weight fusion mechanism based on road and building responses is introduced. The fusion weights are adaptively adjusted according to the degree of support from building structures to complete the fine registration. This invention organically combines high-level semantic information with low-level geometric information, achieving high-precision and robust cross-view geolocation even in the absence of coarse position prior. It effectively solves the problems of easy confusion in pure geometric matching, low accuracy in pure semantic localization, and poor adaptability of fixed fusion methods, making it suitable for global positioning of unmanned systems in complex urban environments.

The invention discloses a water conservancy drawing auditing method and system based on vector analysis and semantic fusion, and relates to the technical field of intelligent water conservancy drawing auditing. The method comprises the steps that a vector file of a hydraulic engineering drawing is analyzed, line segment endpoint coordinates are extracted to construct a graphic information set, and size marking entities are extracted to construct a size marking set; redrawing the vector line segment into a scalar image based on the endpoint coordinates, and outputting a confidence coefficient matrix by using a pre-trained encoder-decoder deep convolutional neural network; in combination with the mapping relation and the confidence coefficient matrix, screening out a target line segment of the target building through a classification decision algorithm; and judging whether the target line segment has an effective size mark or not by utilizing a geometric matching algorithm based on the geometric attribute, and carrying out differential marking on a mark missing object. According to the method, the problems of low identification precision and inaccurate missed mark detection caused by vector information loss due to rasterization and lack of geometric logic verification in the prior art are solved, and high-precision inspection of the water conservancy drawing is realized.

This invention discloses a mine production process monitoring and management system and method based on digital twins, belonging to the field of monitoring and management technology. It includes: a geometric matching module for obtaining matching results based on the geometric information of an underground centralized transfer station; a dust cluster identification module for comparing the dust quality of all internal dust areas in the underground centralized transfer station based on the matching results to obtain the main dust cluster area; a gap screening module for screening gap areas adjacent to the main dust cluster area to obtain the main injection gap; a process dust module for obtaining process dust indicators based on the main dust cluster area and the main injection gap; and a process monitoring module for obtaining production process monitoring results based on the process dust indicators. This invention improves the accuracy and stability of mine production process monitoring and management.

The invention belongs to the technical field of electric power distribution line inspection, and provides an insulator crack-light spot distinguishing method and system.In high-altitude unmanned aerial vehicle inspection, through closed track planning and three-gear exposure synchronous shooting in cooperation with a double-branch detection network, chain type geometric topology constraint and one-dimensional multi-flow graph convolution, the insulator crack-light spot distinguishing method and system are obtained; the method comprises the following steps: firstly, accessing a double-branch detection network by using highlight / dark fringe double-branch lightweight convolution to realize candidate crack detection; according to geometric matching, a candidate frame is aligned with a disc edge, three-exposure local parts are cut, and multi-dimensional features are extracted; after a highlight channel is attenuated by using a saturation mark, assembling features into a chain graph according to a physical sequence, and capturing cross-disk coherent dark fringes; and finally, the full connection layer outputs the crack probability, mirror reflection suppression and real crack suppression are realized, false alarm and missing alarm are effectively reduced, and an accurate basis is provided for a power failure maintenance decision.

The invention discloses a sky-ground collaborative observation method, and relates to the technical field of remote sensing. The method comprises the following steps: firstly, acquiring different coarse / high-resolution remote sensing data, ground measurement and Modis auxiliary data; and adjusting the angle information and the spectral information of the high-resolution remote sensing data to be consistent with those of the coarse-resolution remote sensing data based on the Modis auxiliary data. Extracting a coarse-resolution remote sensing data pixel space range as a template initial position, taking a high-resolution target variable data space resolution as a step length, moving the template, and simulating all geometric positions; and calculating the difference value between the pixel scale target variable aggregation value and the coarse resolution target variable original value during each movement, and determining the real geometric position of the pixel. And finally, in combination with actual ground observation data, calculating a difference value between the scale values of the pixels on the nominal position and the actual position, and quantifying the uncertainty in the sky-ground collaborative observation. According to the method, pixel-level geometric matching among the sky-ground multi-source data in a real sense is realized, the real response range of the reference value extracted from the satellite pixel is ensured, and the limitation of direct comparison in a traditional method is overcome.

Graphical element search technique selection, fuzzy logic selection for anchors and targets, and hierarchical graphical element identification for robotic process automation (RPA) are disclosed. The fuzzy logic selection of anchors and targets may be part of a larger, tiered, or hierarchical process for identifying graphical elements in the UI. When a selector for a UI element is not found with at least a confidence threshold, similar elements potentially corresponding to the selector for a UI element target may be searched based on fuzzy matching of the target and corresponding anchor(s). Geometric matching may also be employed between the target UI element and its respective anchor(s). The combination of fuzzy matching and geometric matching may allow for more flexible and accurate identification of the exact selector with which an RPA robot is attempting to interact.

The application discloses a kind of based on three-dimensional vision's chemical synthesis intelligent operation platform and control method thereof, belong to laboratory automation and intelligent manufacturing field.The platform includes three-dimensional vision acquisition module, point cloud physical property correction module, geometric matching and pose calculation module, motion control and safety constraint optimization module, self-supervision compensation learning module, drift detection and automatic calibration module, safety confidence decision module and mechanical arm execution module, each module is formed closed loop control system by data interface.The application is corrected by constructing refraction compensation model based on Fermat principle, optical deviation caused by transparent vessel and liquid level;Combined with fluid dynamics constraint optimization motion trajectory, prevent liquid splashing and cross contamination;Through self-supervision learning and automatic calibration mechanism, inhibit long-term running in coordinate drift, improve system accuracy and stability.The application can realize high-precision, high robustness and essential safety of automated chemical synthesis operation.

This invention belongs to the field of building construction and digital twin technology, and relates to an intelligent matching method and system based on artificial intelligence. This invention collects construction environment data in real time and identifies risk markers, calculates component deformation using a time-series prediction model, and corrects the virtual components generated by the BIM model for environmental compensation. It employs 3D point cloud registration technology for geometric matching analysis to determine hole alignment errors and interface gaps. Based on a dynamic decision tree model, it dynamically adjusts the installation sequence by comprehensively considering the degree of geometric mismatch, environmental risks, and construction logic. Furthermore, it incrementally updates the model based on actual installation deviations. This invention effectively solves the technical problem of component deformation and installation mismatch caused by environmental factors such as temperature and humidity during construction, achieving dynamic adaptation and precision control of the construction process, improving the matching success rate and construction efficiency of component assembly, and continuously optimizing the system's adaptability and reliability through continuous learning.

This invention provides a three-dimensional quantitative analysis and visualization method and system for aortic diameter expansion velocity based on longitudinal imaging data, applicable to the field of data processing. This application uses baseline and follow-up longitudinal aortic CTA images as the data source. Preprocessing includes spatial resampling and intensity normalization. The aortic lumen is extracted through medical image segmentation, and a voxel mask is generated. Then, the isosurface is reconstructed using the traveling cubes algorithm, and two sets of triangular surface mesh models are obtained through smoothing, cropping, repair, and remeshing. Using the baseline mesh as the source and the follow-up mesh as the target, topologically identical geometric matching is achieved through non-rigid registration. The centerline is extracted, and the point-by-point diameter is calculated. Combined with the follow-up time, three-dimensional quantitative data of the global diameter expansion velocity is obtained. Finally, multi-morphological visualization of expansion velocity and risk stratification is achieved through color mapping, 3D printing, and two-dimensional planar unfolding, constructing a complete three-dimensional quantitative analysis system for the aorta.

The application provides a 4D millimeter wave radar point cloud enhancement and target detection method based on Bayesian fusion, comprising the following steps: step one: projecting the 4D millimeter wave radar point cloud to an image plane and matching the image plane with an image instance segmentation result to obtain an image semantic category and an image segmentation confidence; step two: constructing a conditional probability density model based on the physical characteristics of the radar points; weighting and fusing the image segmentation confidence and the category probability density based on a Bayesian probability framework to obtain a fusion confidence; step three: using the fusion confidence and the semantic category as augmented features to splice the features of the 4D millimeter wave radar point cloud to generate a semantic enhanced point cloud; and using a fusion confidence guided sampling method to sample the semantic enhanced point cloud and inputting the semantic enhanced point cloud into a detection network for training and reasoning. The application solves the technical problems of one-way dependence, shallow geometric matching and insufficient utilization of radar physical characteristics in the existing image and radar point cloud fusion technology.

The invention discloses a vegetable transplanting quality monitoring method based on machine vision, and belongs to the technical field of agricultural information. According to the method, an image is converted to an HSV color space, the weighted color distance of each pixel is calculated based on a pre-stored template plant HSV mean value, and a plant is positioned through adaptive threshold binarization segmentation, morphological optimization and connected component analysis; counting is triggered through a top observation window, inter-frame repeated counting is eliminated through a triangular geometric matching algorithm, and missing planting, falling or normal transplanting is automatically judged according to the adjacent spacing of the plants. According to the invention, illumination adaptability is enhanced, real-time, accurate and automatic monitoring of transplanting quality is realized, and effective technical support is provided for precision agriculture.

This application relates to the field of highway maintenance technology and discloses an intelligent detection and evaluation system and method for highway pavement quality. The invention uses a microsecond-level triggering collaboration and three-dimensional morphological feature surface extraction operator in the underlying hardware to shield non-rigid interference such as water accumulation reflection and exhaust gas. It anchors the inherent rigid cross-section of the tunnel from the multimodal detection data stream, providing an extremely stable sequence of physical feature nodes for subsequent matching. It introduces a slippage cutoff mechanism based on the limit of tire physical static friction to curb blind push displacement drift from the source. At the same time, it reduces the dimension of three-dimensional spatial geometric matching into linear hash sequence comparison, which greatly reduces the on-board computing load.

The invention discloses a loopback detection repositioning method and device based on a laser radar, an inertial measurement unit and large model semantic reasoning, electronic equipment, a storage medium and a program product, and belongs to the technical field of positioning and navigation. The method comprises the following steps: collecting a laser radar point cloud sequence and motion data of an inertial measurement unit in real time; constructing a local geometric feature and a global descriptor based on the point cloud data, and optimizing the matching efficiency in combination with inertial measurement information; after preliminary geometric matching is completed, a large model subjected to multi-modal pre-training is introduced to execute semantic consistency verification and confidence coefficient correction, and a geometric-semantic joint similarity index is generated to suppress false loopback; when effective loopback is detected, geometric constraints, inertial constraints and semantic consistency constraints are jointly introduced into factor graph optimization, and global pose correction is achieved. According to the method, a'geometry-inertia-semantics' three-layer fused closed-loop detection system is constructed, the positioning robustness, continuity and generalization performance in a complex dynamic environment are effectively improved, and high-precision and high-reliability repositioning capability is provided for scenes such as an autonomous mobile robot, intelligent inspection and storage navigation.

This invention discloses a method and apparatus for clustering UAV landslide images for emergency monitoring, belonging to the field of landslide monitoring technology. The method includes: acquiring raw images captured by a camera mounted on a UAV; extracting multimodal features of landslide instances from the raw images, the multimodal features including spatial location features, global semantic embedding features, and local geometric matching features; fusing the multimodal features using an attention-based neural network model to generate instance feature vectors; constructing a graph structure representing the relationships between landslide instances based on the spatial location features, local geometric matching features, and instance feature vectors through an iterative process including a feedback mechanism; extracting connected components from the graph structure, and identifying each connected component as a landslide instance cluster. This invention can accurately associate different image instances belonging to the same landslide, reducing the error rate of cross-frame target matching in complex terrain.

The invention belongs to the technical field of image recognition, and discloses a traditional non-abandoned pattern automatic recognition algorithm. Comprising the following steps: constructing a feature constellation diagram containing feature distribution information and symbiotic relationship strength; obtaining a to-be-recognized image, constructing a base fabric resonance field based on frequency domain characteristics, and constructing a pattern salient field based on color and gradient characteristics; carrying out cooperative enhancement processing on the pattern significant field, carrying out antagonism inhibition by using a base fabric resonance field, introducing residual feedback, and generating a pattern confidence map; performing weighted extraction on a pattern feature vector and a base fabric feature vector based on the pattern confidence map; geometric matching is carried out on the feature vectors and the feature constellation diagram, adaptive weighting is carried out according to symbiotic relationship intensity, and pattern types are determined; and the identification accuracy of the traditional non-abandoned patterns is improved.

The invention discloses an address standardization and space association method based on intelligent semantic analysis, and relates to the technical field of address information processing, and the method comprises the steps: constructing a standard address library comprising space level information and geographic space geometric information according to an administrative space level data set and a house landmark space data set; performing hierarchical analysis on the to-be-processed text address based on the standard address library to obtain target space hierarchical data; and matching a spatial geometric position corresponding to the spatial hierarchy information in the standard address base, and mapping the spatial geometric position to a visual spatial geometric map. According to the method, automatic and batch efficient processing and accurate space matching of the text addresses are realized through standard address library construction, address hierarchical analysis and space geometry matching.

The present disclosure relates to systems, non-transitory computer-readable media, and methods for aligning curves to precise geometric shapes using a geometry matching algorithm. For example, the disclosed systems generate a set of discrete points along a curve depicted in a digital image by sampling the curve spanning between a first corner and a second corner depicted in the digital image. The disclosed systems determine an order for comparing the set of discrete points with a first geometry and a second geometry by comparing a centroid of the curve with a centroid of the first geometry and a centroid of the second geometry. The disclosed systems also determine, according to the order, a first similarity of the curve to the first geometry and a second similarity of the curve to the second geometry. The disclosed systems generate curve segments that align with one of the first geometry or the second geometry.

This invention belongs to the field of electronic information technology and discloses a dynamic guidance method and related apparatus for a large language model based on a conditional generation model. The dynamic guidance method includes: acquiring user prompts; obtaining the activation vector of the user prompts through a specified intervention layer of the large language model; projecting the activation vectors onto pre-extracted semantic bases of each task and identifying the target task to which the user prompts belong through geometric matching; using the target task as input conditions, dynamically generating guidance vectors through a pre-trained conditional generation model, and injecting the guidance vectors into the residual stream of the large language model for dynamic guidance. The technical solution disclosed in this invention can significantly improve the stability and automation level of large language model guidance, while also enhancing generalization ability and robustness.

According to the autonomous examination system for evaluating the marking capability based on the point cloud data, through the data receiving module, the answer configuration module, the geometric matching module and the scoring module, automatic evaluation of the marking quality of the point cloud data is realized, and since a configuration object of a threshold value required by scoring is a point cloud file passing cross-modal matching verification, the marking quality of the point cloud data is greatly improved. As the object configured by the threshold value is dynamically changed, the configured global threshold value and the standard threshold value are also dynamically changed, so that the marking accuracy of the point cloud data is higher, and the judgment accuracy of the marking capability is also higher.

The invention relates to an operation planning method, device and equipment for treating lumbar degenerative diseases through a multi-modal image fusion guided ULIF technology and a medium. The method includes: registering lumbar vertebra CT and MRI images of a patient to generate a fused image data set; respectively reconstructing three-dimensional surface models of a vertebral bony structure, an intervertebral disc and a blood vessel based on the data set; constructing a personalized parameterized finite element model containing the attributes of the heterogeneous material on the basis of the model; performing virtual mechanical simulation on a plurality of groups of preset fusion cage implantation parameters on the model to generate a corresponding biomechanical property index set; on the basis of the index set, Pareto optimal fusion cage implantation parameters are solved through multi-objective optimization; and finally, synthesizing a visual operation planning report containing the optimal parameters. By the adoption of the method, optimization transition from geometric matching to biomechanical simulation can be achieved before an operation, quantitative basis is provided for selection of the model, the position and the angle of the fusion cage, and the method aims at optimizing the mechanical environment of implants so as to reduce the risk of postoperative complications.