89 results about "Model reconstruction" patented technology

The application is a kind of fiber reinforced ceramic matrix composite surface three-dimensional roughness detection method and system, belongs to the technical field of material processing surface quality detection;Method is, first scanning fiber reinforced ceramic matrix composite processing surface to get the original processing surface three-dimensional model, and export point cloud data;Secondly, the coordinate transformation processing is carried out to the original three-dimensional model point cloud data, and the model is put right, the height difference of adjacent point cloud data is taken as the feature, the abnormal data points are screened and removed, the internal hole exposed on the processing surface is removed, the processing surface is reconstructed and the three-dimensional roughness S a Is calculated;Then determine the measurement area;Finally, according to the determined measurement area, the above scanning, hole removal, model reconstruction, S a The calculation work, namely the three-dimensional roughness of the processing surface is determined. The three-dimensional surface roughness value calculated by the application is no longer affected by the hole data, which ensures the effectiveness of the data.

The application discloses a kind of visual enhancement methods for urban road all-weather environment perception, it is related to computer vision and intelligent transportation technical field, steps include: acquisition includes the data set of multiple typical weather and illumination condition;Build physical real cross-domain visual simulation environment, based on training set data, carry out night simulation, dynamic simulation, rain fog simulation, generate the synthetic image with target domain feature;Optimization is carried out to feature, including: the feature fusion based on CutMix and the consistent feature representation based on CORAL Loss;Model adaptation is carried out, from three levels of basic model reconstruction, backbone network light weight, feature fusion structure optimization, systematic adaptation is carried out, improve the perception ability under rain, fog, night complex weather conditions.The method of the application can provide technical support for intelligent perception of urban road under complex weather through real-time deployment and application on edge device.

This disclosure provides a post-processing method for depth estimation, a 3D model reconstruction method, an apparatus, and a device. The specific implementation of the post-processing method for depth estimation includes: acquiring the depth corresponding to multiple pixels in a target image; determining the difference in depth gradients between pixels with neighborhood relationships; wherein pixels with neighborhood relationships include pixels that are close in a target dimension, the target dimension including at least one of the following: color dimension, image position dimension, and depth dimension; minimizing the difference in depth gradients to correct the depth corresponding to each of the multiple pixels.

This invention relates to the field of power engineering identification technology, specifically to a method and system for reconstructing a 3D model of a high-voltage transmission line based on satellite imagery. The method includes: acquiring satellite imagery; identifying tower types; acquiring geographic information data and, in conjunction with the satellite imagery and tower types, performing tower height inversion to obtain the tower height; and reconstructing a 3D model of the high-voltage transmission line using a constructed parametric model library based on the satellite imagery, tower types, and tower height. This solution can efficiently and stably construct a 3D model of a high-voltage transmission line using satellite imagery, thus solving the problem of existing methods relying on expensive data sources and being difficult to apply on a large scale.

The patent application provides a complex environment three-dimensional scene reconstruction method based on a multi-view neural network, and the method mainly comprises two steps: scene information capturing: obtaining scene data from a plurality of views through a high-resolution camera, and ensuring that all details of a complex scene are covered; the captured image is standardized, including color correction and noise removal, to ensure image quality and consistency. And three-dimensional model reconstruction: performing three-dimensional reconstruction on the acquired scene data by using a neural network. Firstly, light sampling is performed on a scene through camera parameters and pose information, and an occupancy grid is created to optimize the sampling efficiency. Thirdly, decomposing the sampling data into feature vectors and inputting the feature vectors into a neural network which is divided into a density network and a color network and is used for generating volume density and color information of the scene; and finally, obtaining a three-dimensional scene model through a volume rendering technology. The method has the advantage that a high-precision three-dimensional model can be quickly generated in a complex environment through an efficient sampling and data processing technology.

This application provides a method for reconstructing an adaptive machining process model for a linear friction welded integral bladed disk, comprising: dividing the blades of the integral bladed disk along the stacking axis into an actual surface region B1, a transition region B2, and a theoretical surface region B3, wherein the upper boundary of the transition region B2 is denoted as S_up, and the lower boundary is denoted as S_down; based on the actual surface region B1, the theoretical blade M_NOM is registered using an iterative nearest-point algorithm, and the blade is moved according to the registration result to obtain the registered and moved blade model M_FIT; based on the upper boundary S_up and the lower boundary S_down of the transition region B2, the adaptive machining transition surface RE_S2 is reconstructed; the actual surface region B1 on M_FIT, the adaptive machining transition surface RE_S2, and the theoretical region B3 on the theoretical blade M_NOM are combined to form an adaptive machining process model for the integral bladed disk blades; each blade is adaptively processed sequentially until all blades on the bladed disk are processed to form the final adaptive machining process model for the integral bladed disk.

The application discloses a method for reconstructing a leading edge model and a trailing edge model of an aero-engine blade. A group of original data points on a blade profile to be machined are first converted into an original theoretical blade profile curve, and then a circle is drawn with the midpoint of the leading edge profile and the trailing edge profile of the original theoretical blade profile curve as the center. Split points are obtained at positions outside the leading edge profile and the trailing edge profile, so as to reconstruct the leading edge profile and the trailing edge profile. All points on the leading edge profile and the trailing edge profile are elongated along a fixed proportion, the digital model reconstruction of the blade profile machining is completed, the machining model with smooth transition is obtained, the leading edge and the trailing edge of the blade profile of the integral blade disc and the integral blade ring can keep uniform and sufficient excess during machining, the leading edge profile and the trailing edge profile of the blade can be effectively ensured, blade damage can be avoided, and the sizes of the leading edge and the trailing edge after machining meet the design requirements.

This invention relates to the technical field of simulation analysis methods for downhole tools in oil and gas fields, and discloses a high-precision dynamic simulation analysis method for full-size drilling and milling tools based on an adaptive FEM-SPH coupled algorithm. The method includes three processes: model establishment, model solving, and performance evaluation. The first process is model establishment, which involves obtaining the geometric model of the drilling and milling tool through model reconstruction technology, and establishing a high-precision dynamic simulation analysis model for the full-size drilling and milling tool based on a wear prediction model and the adaptive SPH-FEM coupled algorithm. The second process is model solving, which involves solving the model after successively changing the drilling pressure and rotational speed, and batch processing the simulation results using Python. The third process is performance evaluation, which evaluates the performance of the drilling and milling tool from three aspects: efficiency, safety, and lifespan. This invention improves simulation accuracy from three aspects: geometric modeling, wear simulation, and solution algorithm, and evaluates the performance of the drilling and milling tool from three aspects: efficiency, safety, and lifespan.

The application provides a 4D person-object interaction generation method based on a text description, which comprises the following steps: in stage one, 3D person-object interaction key frame recovery: a human body motion sequence is obtained through a human body motion model and uniformly down-sampled to extract a key frame; a human body grid is reconstructed through an SMPL-X model for each key frame to extract a vertex position and form a human body point cloud; an object position anchoring network takes the human body point cloud, an object template point cloud and a text prompt as input to predict an object position and generate a 3D person-object interaction key frame; in stage two, 4D person-object interaction sequence generation: a contact perception diffusion model is constructed to take a sparse 3D person-object interaction key frame as input, extract a conditional signal through a contact perception encoder; based on the conditional signal, the 3D person-object interaction key frame is subjected to time series interpolation through the contact perception diffusion model to generate a time series coherent dense 4D person-object interaction sequence. The application realizes natural and realistic 4D person-object interaction synthesis of an unseen object.

The application provides a three-dimensional scanning point cloud reconstruction method, system, device and medium for complex workpieces, comprising: processing initial scanning data of a complex workpiece to obtain a target point-normal set; generating a candidate scanning viewpoint set according to the set; performing visibility evaluation, registrability evaluation and robot executability screening on the viewpoint set to obtain a scanning viewpoint sequence; sequentially moving to each scanning pose according to the sequence, and controlling a three-dimensional scanning sensor installed at the end of the robot to collect multiple frames of local point clouds, and obtaining overlapping sub-clouds according to the point clouds combined with a nearest neighbor constraint strategy; performing progressive continuous fine registration on adjacent frame point clouds according to the overlapping sub-clouds to obtain fine registration poses of each frame of point clouds relative to a target frame; and fusing multiple frames of local point clouds according to the poses to obtain fused point clouds; and the above method improves the completeness, splicing accuracy and automation degree of three-dimensional point cloud model reconstruction of complex workpieces.

This disclosure provides a method, apparatus, device, and storage medium for three-dimensional model reconstruction, relating to the field of artificial intelligence technology. In some embodiments of this disclosure, a measured depth map and a normalized depth map corresponding to a three-dimensional scene are obtained; the measured depth map and the normalized depth map are aligned to obtain a set of effective pixels; based on the set of effective pixels, a target scale factor is determined according to an iterative weighted least squares algorithm and a robust loss function; the coordinates of three-dimensional points in the normalized three-dimensional model are reconstructed according to the target scale factor to obtain a three-dimensional model corresponding to the three-dimensional scene. This disclosure, by aligning the measured depth map and the normalized depth map and combining the target scale factor, achieves automatic mapping from the normalized three-dimensional model to the real physical scale, improving the accuracy of three-dimensional model reconstruction.

The application discloses a kind of human target infrared three-dimensional model fast reconstruction method, belong to model reconstruction algorithm field.The method is first based on infrared physics to the heat transfer process on human surface is physically modeled, and utilizes the principle of graphics to generate the human infrared three-dimensional feature image of target human body under different conditions.Functions are 1) based on measured data, establish the infrared radiation database of simulation target environment.2) the heat transfer process on human surface is physically modeled using finite difference method, realizes three-dimensional human infrared model drawing.3) using computer iterative calculation generates the infrared feature image of human target in different time, place, climate condition.4) based on the principle of Planck spectrum energy distribution, long-wave infrared simulation image of human body is converted into mid-wave simulation image.Through hash algorithm, the similarity and reliability of human infrared simulation imaging result are evaluated, and the simulation similarity is higher than 70%.

The application discloses a multi-parameter vital sign monitoring and early warning system based on an intelligent ambulance, and relates to the technical field of signal alarm devices.The system comprises a task receiving and parameter determining module, a model reconstruction and compression module, and a model issuing and monitoring and early warning module.The specific steps include receiving first-aid task information to dynamically determine a list of target vital sign parameters;reconstructing and compressing a cloud-integrated early warning model based on the list to generate a lightweight early warning model;finally, issuing the lightweight model to the target ambulance to perform monitoring and early warning, and adopting a local and cloud collaborative early warning mechanism based on confidence.The application solves the technical problems of insufficient early warning accuracy caused by the lack of task adaptability, the difficulty in deploying complex models due to the limitation of computing resources, and the impact of the local and cloud alarm logic on the collaborative efficiency, and realizes accurate, real-time and reliable early warning of vital signs in a mobile first-aid scene with limited resources.

This invention proposes a driving scene reconstruction method and apparatus based on online learning to generate priors, relating to the field of 3D scene reconstruction technology. The method includes: using a video of the original trajectory, training a 4DGS model after depth estimation initialization to obtain a reconstructed 4DGS model; after the reconstruction model converges, sampling the 4DGS reconstruction model using a new trajectory to obtain a rendered video under the new trajectory; reprojecting structured information onto the image under the new trajectory using projection transformation, generating a corresponding video using the structured information and constraints from the first frame image, and training a generative model using the rendered image of the new trajectory as supervision; after the generative model converges, using the generated video under the new trajectory to further optimize the reconstruction model, iterating in this way until both models fully converge, thus achieving driving scene reconstruction.

The application relates to the technical field of model processing, and provides a CAD model reconstruction method, a system, a device and a medium.The method comprises the following steps: acquiring an original CAD model, an original grid and a deformed grid; the deformed grid is partitioned according to the original CAD model and the original grid, so that a plurality of feature regions are obtained; parameter information of grid vertices in each feature region is acquired, so that a parameter information set is obtained; the grid vertices are converted into structured point clouds according to the parameter information set and spatial coordinates of the grid vertices; an uncut surface is generated based on the structured point clouds; the uncut surface is cut, so that a cut surface is obtained; and the cut surface is spliced, so that a reconstructed solid CAD model is generated.

The embodiment of the application provides a complex stratum cave volume calculation method and related equipment, and belongs to the technical field of geological survey. The method comprises the following steps: obtaining a geological profile and a drilling columnar graph; performing information extraction on the drilling columnar graph to obtain drilling information of each drilling point; performing structured processing on the drilling information of all drilling point positions to obtain an original stratum distribution table; performing elevation supplement on the original stratum distribution table according to the geological profile to obtain a standard stratum distribution table; based on the standard stratum distribution table, performing data splitting on each stratum as a derived unit to obtain a plurality of stratum derived tables; performing three-dimensional model reconstruction on each stratum derived table to obtain a stratum interface entity model of each stratum; performing target detection on all stratum interface entity models according to a preset cave label to obtain a target cave model; and performing face domain attribute analysis on the target cave model to obtain a target cave volume. The application can improve the accuracy of cave volume calculation.

This invention discloses an internal diameter detection method and system based on multi-sensor fusion and adaptive compensation, belonging to the field of mechanical inspection technology. The method includes steps such as: benchmark calibration, spiral scanning, multi-sensor synchronous acquisition, dynamic error compensation, coordinate transformation, three-dimensional internal cavity model reconstruction, geometric tolerance calculation and evaluation, dynamic calibration and error update, and equipment status monitoring. The system includes a workbench, positioning mechanism, lifting actuator, lifting pan-tilt head, internal diameter detection head, rotary drive mechanism, angle encoder, environmental sensor, and controller. It achieves full coverage detection of the internal cavity through multi-sensor fusion and can simultaneously evaluate multiple geometric tolerances; it eliminates system errors and environmental interference through a three-level compensation mechanism to ensure long-term detection accuracy; it visualizes the internal cavity shape through three-dimensional model reconstruction; and it enables predictive maintenance through equipment status monitoring.

This application discloses a method, system, apparatus, storage medium, and electronic device for three-dimensional model reconstruction. Relating to the field of image processing, the method includes: acquiring a set of voxel data collected by a scanning device in a target to be reconstructed during the current scanning phase; the set of voxel data includes voxel data after voxelizing the set of location points collected by the scanning device; determining a first isosurface based on the voxel data set, and determining an incremental voxel data set based on the first isosurface; the first isosurface representing the internal structural surfaces of the target to be reconstructed, and the incremental voxel data set including voxels within the first isosurface; fusing the incremental voxel data set with historical incremental voxel data sets corresponding to historical scanning phases to obtain a target voxel data set; and reconstructing the internal structure of the target to be reconstructed based on the target voxel data set. This application solves the problem of low reconstruction accuracy in related technologies that perform a one-time full surface reconstruction of the target to be reconstructed.

The application provides a small animal sciatic nerve magnetic resonance coil and an imaging method, and the method comprises the following steps: filling a high-concentration sugar syrup solution at the bottom of the magnetic resonance coil and performing constant temperature control; placing a small animal in a prone position in the high-concentration sugar syrup solution and discharging air in the gap between the lower limbs and below the abdomen of the small animal; after performing magnetic resonance positioning scanning to determine the main sequence scanning range, starting the main sequence scanning, suppressing the flow tissue signal through a pre-pulse combination, exciting the tissue by using a radio frequency pulse with preset parameters, acquiring echo signals by combining a fast spin echo technology and filling K space, and obtaining scanning original data; performing preprocessing on the scanning original data and accurate segmentation of the sciatic nerve and spinal nerve to obtain a segmentation result; and completing three-dimensional model reconstruction, optimization and multi-modal fusion visualization based on the segmentation result to obtain an image scanning result. The application is suitable for small animals, is accurate in imaging, can effectively suppress various kinds of artifacts, and can be widely applied to sciatic nerve imaging scenarios of small animals.

The application discloses a distributed optimization model reconstruction method based on a complex chemical process and a related device, relates to the technical field of chemical system engineering and numerical optimization calculation, and comprises the following steps: acquiring system structure parameters, flow connection parameters and process constraint parameters of a target chemical process system; decoupling and dividing the target chemical process system into a plurality of local subsystems according to a coupling relationship between material flow and energy flow; determining a selected distributed optimization algorithm; performing mathematical reconstruction on coupling boundary variables and consistency constraints between the local subsystems according to a corresponding model reconstruction strategy matched with the architecture type of the distributed optimization algorithm, to obtain a reconstructed mathematical model; performing distributed parallel iteration solving on the reconstructed mathematical model, and outputting a distributed optimization result. The application can improve the optimization efficiency and effect of the chemical process system, thereby reducing the operation cost of the chemical process system and improving the operation performance of the chemical process system.

The application discloses a social interaction method and system based on GIS space-time base and digital twin simulation, relates to the field of digital twin and GIS fusion, and comprises the following steps: collecting multi-source heterogeneous data of a target geographic area through sparse sensing nodes, and encoding the data into implicit space-time tokens by using a Bit space-time variational autoencoder; inputting the tokens into a generative world model based on a diffusion transformer as a unique condition, and reconstructing a four-dimensional space-time field; converting the four-dimensional space-time field into a three-dimensional Gaussian sputtering representation, and rendering the representation into a two-dimensional view in real time by using a progressive resolution strategy; updating Gaussian sputtering parameters asynchronously through a time series prediction network in response to newly-arrived sparse sensing data; constructing a social interaction layer on the rendered view, and supporting multi-user distributed access; and predicting a future video frame sequence by using a random mask strategy in response to an interaction instruction. The application realizes the construction of a full-scale, high-fidelity and interactive digital twin world from a macro geographic space to a micro device state.

The application provides a heat dissipation structure design method of an intelligent processor, comprising the following steps: a model reconstruction step, combining a design process of a structure design software and a thermal simulation software, reconstructing models according to modeling results of the two design processes, and obtaining temperature results of the two design processes; an error calculation step, the thermal simulation software calculating error parameters of the two design processes according to the temperature results of the two design processes; and an equivalent transformation step, the structure design software performing equivalent transformation design according to the error parameters. The application also provides a heat dissipation structure design device of an intelligent processor, a storage medium and an electronic device. Therefore, the application fully utilizes the advantages of the two heat dissipation structure design methods, simplifies the process, reduces the design cost, and improves the design efficiency.

This invention discloses a method and apparatus for compressing and inheriting ancestor model knowledge into descendant models through optimal transmission. The method includes: constructing an ancestor model containing multiple Transformer network layers and obtaining a unified feature representation for each layer of the ancestor model; constructing a difference metric between each layer of the ancestor model and candidate learning genes; modeling the cross-layer knowledge mapping process as an entropy-regularized optimal transmission problem; solving the entropy-regularized optimal transmission problem to obtain a cross-layer knowledge transmission matrix; performing cross-layer weighted fusion of ancestor model parameters based on the cross-layer knowledge transmission matrix to construct shared learning gene parameters; constructing descendant models of different scales based on the learning gene parameters; and applying the constructed descendant models to the training and inference of downstream tasks. This invention can complete structured knowledge inheritance and model reconstruction without accessing the original training data of the ancestor model, and has good deployability and generalization ability.

The present application relates to the technical field of image reconstruction, in particular to a liver tumor interventional preoperative blood vessel three-dimensional model reconstruction system, which can realize the following steps through the mutual cooperation between multiple modules: obtaining a liver CTA three-dimensional matrix of a target patient before liver tumor intervention; determining the blood vessel confidence corresponding to each voxel in the tumor peripheral region and the normal liver region; determining the gradient distortion degree corresponding to each voxel in the tumor peripheral region; determining the effective blood vessel proportion in the tumor peripheral region; performing adaptive blood vessel fracture completion screening processing according to the gradient distortion degree and the blood vessel confidence to obtain the tumor peripheral blood vessel region; and performing blood vessel reconstruction based on the effective blood vessel proportion in the tumor peripheral region and the gray value corresponding to the voxel in the tumor peripheral blood vessel region. The present application realizes the identification of the tumor peripheral blood vessel region and improves the accuracy of identification, thereby improving the rationality of liver tumor interventional preoperative blood vessel three-dimensional model reconstruction.

The application discloses a kind of based on dynamic threshold and double monitoring strategy fault detection method and system, it is related to the field of fault detection, specific steps are as follows: based on historical process monitoring data, build comprehensive detection index and initial threshold;Initial threshold is adaptively dynamically updated based on dynamic threshold real-time updating mechanism, generate the first threshold for abnormality determination;According to the abnormality cumulative statistics of comprehensive detection index, and set the second threshold for abnormality determination;Acquire the process monitoring data to be processed, and calculate the comprehensive detection index of current time, in the retrieval path of pre-construction, according to the first threshold and the second threshold parallel detection is carried out, and abnormality determination result is obtained.The application is adjusted mechanism based on dynamic threshold and double monitoring strategy, under the condition that without additional prior knowledge and complex model reconstruction, multiple working conditions, multiple fault type industrial process is efficiently and accurately detected.

This invention discloses a high-precision method for reconstructing a directed distance field model in the field of solid rocket motor burnout calculation technology. The method includes: obtaining the directed distance field function based on a fixed mesh; reconstructing initial isosurfaces / lines using linear interpolation based on the type and spatial dimension of the fixed mesh, and marking isosurfaces / lines that make the directed distance field function zero; and obtaining and classifying the mesh nodes of the mesh cells intersecting the isosurfaces / lines obtained by searching using linear interpolation, resulting in directly adjacent nodes, a set of directly adjacent nodes, and a set of non-adjacent nodes. This invention constructs a reference base circle and uses a local optimization method to find isosurfaces / lines that are tangent to as many base circles as possible, ensuring that the directed distance field calculated based on the reconstructed isosurfaces / lines is consistent with the input directed distance field, avoiding non-physical offsets of the isosurfaces / lines, thereby eliminating systematic errors and improving calculation accuracy.