342 results about "Real-time rendering" patented technology

The invention discloses a three-dimensional scene reconstruction method and device based on large model geometric prior, and a medium, and aims to solve the problems that a conventional 3DGS is liable to have artifacts and detail loss in geometric discontinuity, data redundancy and illumination variation scenes, and predicts a dense depth map and a normal map from a monocular image by using a pre-trained large model. The position and form of the Gaussian kernel are constrained as additional geometric priori; a primitive adjustment strategy based on kernel density estimation is introduced in the training stage, small Gaussian primitives with similar structures and adjacent spaces are combined into a large Gaussian primitive, the rendering quality is kept, redundancy is reduced, and the volume of the model is reduced; an exposure coefficient is adaptively estimated for each input image, an exposure compensation image loss function is constructed, and floating artifacts caused by illumination differences at shooting moments are eliminated. Experiments show that compared with the prior art, the method improves the three-dimensional reconstruction precision and real-time rendering quality of complex illumination and less-texture areas in a public data set and an unmanned aerial vehicle aerial photography scene.

The invention discloses a Gaussian model construction method and device, a program product and a storage medium, and relates to the technical field of computer vision. The method comprises the following steps: acquiring multi-view images, and obtaining a three-dimensional Gaussian element corresponding to each multi-view image; establishing a hierarchical data structure; selecting a training view angle to obtain a trained three-dimensional Gaussian element; calculating a stability score of each trained three-dimensional Gaussian element, and performing hierarchical construction operation according to each stability score to obtain a plurality of space clusters; calculating a representative parent cluster; performing rasterization processing to obtain a rendering result; calculating rendering loss, and adding the representative parent cluster into the hierarchical data structure to obtain a target hierarchical data structure; and repeatedly executing the training operation and the hierarchical construction operation until the rendering loss is lower than a preset loss threshold value, and outputting a three-dimensional Gaussian scene model containing the target hierarchical data structure. By implementing the technical scheme provided by the invention, the performance pressure faced by real-time rendering can be reduced.

The invention relates to the technical field of image processing, and provides an indoor three-dimensional image intelligent rendering method based on image processing. Comprising the following steps of indoor multi-view image acquisition, image preprocessing, intelligent analysis of indoor scene elements, construction of an indoor three-dimensional initial model with attributes, intelligent generation of rendering parameters, adaptive LOD real-time rendering and intelligent interaction optimization. According to the intelligent interaction optimization system, through deep fusion of natural language processing, user preference learning and real-time rendering technologies, a set of efficient, visual and personalized virtual scene rendering interaction process is constructed. According to the method, the problems that traditional graphic software is complex in operation, high in learning cost and low in debugging efficiency are solved, and the user satisfaction and creation efficiency are remarkably improved through an intelligent recommendation and rapid iteration mechanism.

The invention relates to the field of computer software, and provides an AI-based low-code visual component dynamic generation method and system. The method comprises the following steps: receiving a component demand description input by a user, and converting the component demand description into structured cue word data to obtain an initial cue word sequence; retrieving matched template data and grammatical rule data from a component knowledge base according to the initial cue word sequence, and fusing the template data, the grammatical rule data and the initial cue word sequence to obtain an enhanced cue word sequence; performing semantic analysis and code generation processing on the enhanced cue word sequence through a large language model to obtain a component source code; and performing real-time rendering processing on the component source code through a component rendering engine, and generating an interactive component instance in a preview interface to obtain reusable component resources. According to the method, the code quality consistency is ensured, and the development efficiency is improved.

The invention relates to the technical field of forest fire monitoring and power grid safety prevention and control, and discloses a fire scene spreading time sequence reconstruction method based on Doppler weather radar mountain fire echoes, which comprises the following steps: acquiring original data of the Doppler weather radar mountain fire echoes and carrying out quality control processing to obtain radar echo data; performing automatic segmentation of a smoke plume area on radar echo data by adopting an optimized Otsu-Unet network model, and extracting a centroid coordinate of a mountain fire echo area; a Kalman filtering algorithm is used and combined with external three-dimensional wind field data to predict a future motion trend of centroid coordinates of a forest fire echo region so as to reconstruct a continuous fire scene spreading time sequence voxel sequence; performing 3DTiles slicing processing on the fire scene spreading time sequence voxel sequence, and performing real-time rendering display on a power grid monitoring platform by using a WebGL technology; and obtaining an external expansion boundary of fire scene spreading based on the fire scene spreading time sequence voxel sequence, and calculating the distance between the external expansion boundary and the power transmission line GIS data in real time. The fire prevention and control capability of the power transmission line of the power grid is improved.

The invention discloses a scene generation method and device based on multi-source GIS data fusion, and relates to the technical field of digital twinning and programmed generation. The method comprises the following steps: acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource library; a structured resource library is constructed, semantic parameters are added to the three-dimensional model in the structured resource library through the configuration file, and three-dimensional model resources with structured semantics are constructed; on the basis of the three-dimensional model resources with structured semantics and GIS data in a geographic information resource library, building and road generation and terrain processing are carried out in a programmed modeling engine through a configuration file, and scene data are generated; and importing the generated scene data into a real-time rendering engine, carrying out dynamic environment interaction and biocenosis simulation, and generating a city scene. The problems that in the prior art, an urban three-dimensional modeling method is low in efficiency and insufficient in environment interaction reality sense are solved.

The invention relates to the technical field of digital twinning, in particular to a real-time interactive visualization system and method based on digital twinning, and the system comprises a data collection module which is used for configuring a multi-protocol adapter deployed on a physical equipment layer to collect the operation data of physical equipment, and transmitting the operation data of the physical equipment to an edge computing node. According to the method, the general data adapter is used for driving the three-dimensional rendering engine to execute real-time rendering based on physics, so that the virtual scene can dynamically reflect details such as illumination change and material aging of a physical entity, and high-fidelity visual feedback is provided. According to the method, a closed-loop control mechanism is established in combination with a multi-modal interaction framework, pure data monitoring is converted into operable bidirectional interaction, operation of a user on a visual interface can be converted into feedback instructions in real time to reversely adjust physical equipment, and the response efficiency of remote operation and maintenance is improved.

The invention relates to the technical field of robot perception and data generation technologies, in particular to a first-view-angle drilling method and device based on memory enhancement and a storage medium, and the method comprises the steps: extracting a plurality of target frames from a historical video stream stored in a space intelligent machine, obtaining a plurality of memory elements based on the plurality of target frames, performing three-dimensional reconstruction on each target frame, constructing a target world model, planning a first visual angle track of the target robot in the target world model, performing imaging simulation on the target robot, performing real-time rendering on a first visual angle video of the target robot, and performing real-time rendering on a second visual angle video of the target robot based on the same time axis and action script as the first visual angle video. A third-person video corresponding to the space intelligent machine is generated, and a drilling result is output; according to the method, the long-term memory data of the space intelligent machine is ingeniously used, high-quality drilling data can be quickly generated, the reliability of the first-view video is improved, and a reliable basis is provided for training and testing of a robot algorithm.

The invention discloses an AI digital human real-time rendering method based on GPU acceleration, and relates to the technical field of digital human rendering, and the method comprises the steps: S1, constructing a multi-GPU hardware feature and load monitoring module, and collecting hardware feature parameters and current load states of each GPU participating in collaborative rendering in real time; according to the method, by constructing the multi-GPU hardware feature and load monitoring module and combining a dynamic task allocation algorithm, dynamic allocation of the rendering sub-tasks to the optimal GPU is achieved, the problems of GPU performance bottleneck and resource idleness caused by traditional fixed task allocation are solved, the multi-GPU collaborative rendering efficiency is improved, and the multi-GPU collaborative rendering efficiency is improved. A scene complexity analysis module and a self-adaptive resource allocation algorithm are built, the proportion of GPU resources between an AI driving module and a rendering module is dynamically adjusted according to scene complexity, the AI digital human interaction naturalness is improved in a simple scene, the rendering frame rate and the picture quality are guaranteed in a complex scene, and the method is suitable for being used in a large-scale scene. And finally, dual optimization of rendering efficiency and picture quality in AI digital human real-time rendering is realized.

The invention belongs to the field of automatic driving, and particularly relates to a multi-scene fusion automatic driving analog simulation test method and system, and the method comprises the steps: obtaining multi-source heterogeneous driving data, carrying out the preprocessing and fusion of the multi-source heterogeneous driving data, and obtaining standardized scene data; based on the standardized scene data, constructing a basic simulation scene by adopting a photo-level real-time rendering technology, and optimizing and constructing the basic simulation scene into a three-dimensional scene model by adopting a 3D Gaussian splashing technology; based on a preset scene generation rule and through a 4D space-time modeling algorithm, performing expansion and variation on the three-dimensional scene model, and automatically generating a multi-dimensional test scene set; executing a simulation test corresponding to the multi-dimensional test scene set, and collecting test result data of the automatic driving system; and establishing a scene optimization and algorithm iteration closed-loop mechanism based on the test result data, and dynamically optimizing the multi-dimensional test scene set and the automatic driving system algorithm.

The invention discloses an image super-resolution edge reconstruction method based on morphology and coverage rate perception. The method comprises the following steps: dividing an image into an edge region and a non-edge region based on depth and brightness information; anti-aliasing edge detail recovery is realized through coverage rate perceived depth mode classification and a table look-up filter; generating direction features by using the brightness gradient and keeping details through a table look-up filter; and combining the reconstruction results of the current frame and the historical frame to ensure the image stability in a dynamic scene. The invention also provides an image super-resolution edge reconstruction system. According to the method provided by the invention, the quality of an object edge and a newly appearing pixel region can be remarkably improved, sawtooth artifacts are effectively inhibited, and relatively high time stability is kept in a dynamic scene, so that the requirements of real-time rendering and visual enhancement application are met.

The invention provides a three-dimensional science popularization animation generation method and system based on an AIGC technology, and relates to the technical field of animation generation. The method comprises the following steps: firstly, fusing structured knowledge such as an anatomy term library, a diagnosis and treatment process guide and a drug specification with electronic medical records and CT / MRI volume data, and generating a split script and a medical term list through a large language model; then reconstructing an organ or focus grid by using a script-image combined driven text-three-dimensional diffusion model, and executing anatomical consistency correction according to the medical knowledge graph; automatically arranging a time axis, a camera track and a skeleton action, and synchronously synthesizing multilingual parastyle subtitles; and finally, frame-level rendering and compression packaging are completed in a real-time rendering engine, and efficient, professional and iterable medical animation generation is realized by means of an expert feedback closed-loop fine tuning model.

The invention discloses a localization-adaptive track point visual rendering method based on partitioning and hierarchical caching, and the method comprises the steps: S1, inputting a track data stream, carrying out the reading according to the source type of the data stream, and carrying out the data partitioning and LOD processing of the track point data; s2, performing data analysis and rendering on the data; s3, carrying out batch processing on GPU vertexes, according to the state of the circular buffer, if the vertexes are writable, incrementally writing vertex data, and if Wrap is needed, carrying out loopback writing and updating a pointer; and S4, finally, the descriptors and the styles are updated, rendering increment drawing is carried out on the track points, and visual display is carried out. The method supports real-time rendering and playback of ten-million-level track points, has visual and instantiated rendering of window dynamic loading, space-time partitioning, hierarchical caching and WebWorker decoupling calculation, and improves the space-time data rendering capability of the domestic autonomous controllable field by ten thousand times; and the user experience is greatly improved by a non-perceptual interaction and transition smoothing algorithm.

The invention discloses a 3D RS and pyramid model fused optical flow calculation method for real-time rendering frame insertion, and the method comprises the steps: 1, obtaining the information of each texture in a rendering frame, so as to construct a texture pyramid; step 2, selecting a grey-scale map corresponding to the lowest resolution to estimate an optical flow initial value; step 3, generating an optimal up-sampling optical flow according to the optical flow initial value, and updating the optimal up-sampling optical flow in a neighborhood to obtain an updated optical flow; 4, repeating the step 3 to obtain the up-sampling optical flow with the highest resolution; 5, performing fine adjustment on the up-sampling optical flow under the highest resolution; step 6, calculating motion vectors of target pixels in the previous rendering frame and the current rendering frame based on the rendering pipeline, and generating an optimal motion vector in combination with the up-sampling optical flow after fine tuning of the target pixel block; and 7, reconstructing a final intermediate frame image. The invention also provides an optical flow calculation system. The method provided by the invention can effectively solve the problem of wrong sampling of real-time rendering frame insertion.

The invention relates to the technical field of three-dimensional modeling, and provides a lightweight hierarchical digital twin modeling and rendering method for a large-scale scene. The method comprises the following steps: analyzing the surface geometry of an original high-precision three-dimensional model, calculating and extracting curvature characteristics of the high-precision three-dimensional model corresponding to each local area, and constructing curvature response distribution through the curvature characteristics; screening and judging each model region corresponding to the original high-precision three-dimensional model based on the obtained curvature response distribution and the dynamic curvature threshold, and obtaining a second key detail region; generating an independent multi-resolution detail enhancement data block for the second key detail area, generating a plurality of data versions with gradually decreased detail degrees for the same second key detail area, and forming a multi-resolution detail library capable of being directly indexed; in the real-time rendering stage, seamless fusion rendering is completed according to the detail data blocks and the basic main body model, and unification of overall scene light weight and local detail high fidelity is achieved.

The invention discloses a structural semantic guided bridge point cloud Gaussian splash modeling and rendering method, and belongs to the technical field of computer vision and bridge engineering digitization. The method comprises the steps of collecting a multi-view image of a bridge, and obtaining a sparse point cloud and a camera pose through a motion recovery structure technology; dense point clouds are generated, deep learning semantic segmentation is carried out, and key components such as piers, bridge floors and cables are identified; mapping the semantic tag back to the sparse point cloud; performing differential Gaussian element initialization on the sparse point cloud based on semantic tags, and endowing different parts with initial shapes conforming to geometric characteristics of the different parts; a geometric constraint loss function is introduced in the optimization process for joint optimization; and finally, outputting a Gaussian splash model of a high-fidelity and accurate geometric structure. According to the method, the problems of model redundancy and geometric distortion when a traditional Gaussian splashing technology is used for processing a bridge scene are solved, and lightweight and high-precision reconstruction and real-time rendering of the bridge model are realized.

The invention discloses a three-dimensional model dynamic rendering method and a storage medium, and the method comprises the steps: obtaining three-dimensional data of a target scene, obtaining a plurality of voxels, and enabling each voxel to correspond to a plurality of voxel attributes and an initial comprehensive displacement vector; in any voxel in any time step, obtaining a comprehensive displacement vector of the voxel in the current time step according to a preset physical deformation rule and / or animation curve deformation rule, and performing coordinate transformation on the voxel; after the coordinate transformation, updating the voxel attribute of the voxel corresponding to the current time step length; and performing real-time three-dimensional rendering on the target scene according to the vertex normal of each voxel at the current time step length and the voxel attribute. The physical consistency is ensured in the dynamic deformation process by synthesizing the displacement vector; by updating the voxel attributes, the voxel attributes are kept in smooth transition after deformation; and high-quality real-time rendering of the target scene is realized by recalculating the deformed vertex normal and combining the illumination model.

The invention discloses a material database system with enhanced artificial intelligence. The material database system comprises an interactive user interface module, a data management and service module, a visual processing module and an AI analysis and intelligent dialogue module. Structured and unstructured material data are managed in a centralized manner through a unified database and cloud storage, and a standardized data center is constructed; by integrating a chemical structure rendering library and a 3D graphic engine, real-time rendering from a CIF file to an interactive 3D model is realized, and a dynamically generated XRD spectrogram is displayed through an interactive chart library, so that the interpretability of data is greatly improved; by introducing an AI application development framework and an LLM, the system is endowed with intelligent analysis ability exceeding simple data retrieval, and professional material interpretation can be provided; by integrating the RAG system, a traditional search mode is upgraded into an intelligent dialogue mode, so that a user can explore material data like a dialogue with an expert.

The invention relates to the technical field of computer graphics and human-computer interaction, in particular to an immersive literature interaction system based on WebGL and 3DGS, which comprises a scene reconstruction and data processing module based on 3DGS, a 3DGS real-time rendering engine based on WebGL and an interaction narration module fused with a 3DGS scene. The 3D GS-based scene reconstruction and data processing module is used for reconstructing a high-precision 3D scene from multi-source data and executing lightweight optimization; the WebGL-based 3DGS real-time rendering engine is used for loading and rendering a 3D Gaussian splash model at a browser end, supporting dynamic illumination, texture and viewpoint switching and realizing low-delay interaction; and the interactive narrative module fusing the 3DGS scene associates literature content through AI driving logic, triggers scene visualization response and integrates a role dialogue system to realize immersive narrative experience. The module realizes seamless fusion of literature interaction logic and a 3DGS scene.

The invention discloses a semantic-guided lightweight three-dimensional reconstruction method based on three-dimensional Gaussian, and the method comprises the steps: inputting a multi-view two-dimensional image, and extracting two-dimensional semantic features; projecting the two-dimensional semantic features to the three-dimensional Gaussian primitives; learning three-dimensional semantic distribution in a knowledge distillation mode, and predicting three-dimensional semantic embedding features; dividing a large scene into a plurality of sub-regions, and performing divide-and-conquer three-dimensional Gaussian optimization and semantic learning; performing semantic-guided lightweight modeling on the three-dimensional Gaussian primitives in each sub-region; executing sub-scene fusion, and performing quantitative compression on the three-dimensional data in the fused scene; and a rasterization renderer is used to carry out efficient real-time rendering under any visual angle, and a lightweight large-scene three-dimensional reconstruction result is output. According to the method provided by the invention, the real-time rendering performance of a complex scene can be remarkably improved, the model volume is effectively compressed, meanwhile, the model training process is accelerated, and the method is suitable for efficient three-dimensional visual reconstruction tasks.

The invention discloses a 3D Gaussian sputtering human avatar reconstruction method based on attitude learning, and the method comprises the steps: initializing a scene through 3D Gaussian primitives for a monocular dynamic video, enabling the Gaussian primitives to be embedded into a triangular gravity center coordinate system under the drive of an SMPL-X grid, and achieving the one-to-one mapping from a standard space to an attitude space; a multi-layer perceptron is utilized to predict attitude-related offset, non-rigid fine adjustment is performed on Gaussian positions and shapes, and primitives are periodically added and deleted through adaptive density control to suppress redundancy. And finally, combining the Gaussian sets corresponding to the skeleton motion and the texture details, and outputting a human avatar model which can be rendered in real time. According to the method, multi-view collection is not needed, training and reasoning resource consumption is low, and consumption-level hardware deployment is facilitated.

The invention relates to a rapid rendering method for an urban Beidou cubic grid. The method comprises the following steps: firstly, acquiring urban low-altitude geographic space data; constructing a multi-level grid system according to conflict detection precision and response requirements by using a Beidou cubic grid subdivision technology, and associating codes and airspace attributes with grid units; unloading the geometric generation task to a GPU for parallel calculation, and generating rendering adaptive grid geometric data; then developing a three-dimensional rapid rendering algorithm to realize real-time rendering of any region and precision grid; subsequently establishing a rendering result and airspace management cooperation mechanism to assist route planning, management and control and zoning; and finally, an algorithm optimization module is linked. The invention aims to solve the problems that the existing ground object 3D rendering adaptive city low-altitude field is insufficient, the conflict detection precision is not matched, the real-time performance is poor and the existing ground object 3D rendering adaptive city low-altitude field is disjointed with a low-altitude management service.

The invention relates to the technical field of digital content production, and discloses an interactive digital content production system based on a Transform architecture. According to the system, text, image and audio data of original digital content are acquired through a content feature extraction module, cross-modal feature alignment is performed by using a multi-head attention mechanism, and content feature tensors with space-time relevance are generated; the dynamic weight distribution module calculates relative importance scores of different modal features based on the tensor, and adopts a gating mechanism to perform dynamic weight fusion to form content semantic enhancement representation; the interaction intention analysis module performs space-time coding matching on the enhanced representation and the user operation instruction stream, and analyzes an intention distribution matrix of user operation on the content dimension; a hierarchical decoding generation module constructs a multi-scale content generation path in a Transform decoder according to the intention distribution matrix; and the real-time rendering engine module loads implicit representation output by the path, so that efficient and intelligent digital content creation is realized.

The invention relates to the technical field of distance education, and discloses a multi-modal interaction-driven distance education immersive scene construction method and system, and the method comprises the steps: configuring a scene resource scheduler of a virtual classroom according to the real-time interaction intention of an education participant; extracting a scene resource scheduling list and a dynamic content parameter set in the scene resource scheduler to construct a dynamic scene map of the virtual classroom in combination with the real-time interaction intention; establishing a real-time rendering task queue by using an immersive narrative flow generated by the dynamic scene map and an interactive response rule set, and rendering an immersive teaching plot of a virtual classroom; and based on the dynamic scene map, generating a distance education immersive scene corresponding to the education participant by rendering the task queue and the immersive teaching plot in real time. According to the invention, sudden demands in the teaching process can be flexibly responded, and the situation adaptability of a remote education scene is improved.

The invention relates to the technical field of computer graphics, sensor simulation and real-time rendering, in particular to a panoramic laser radar simulation method and system based on a single GPU calculation shader, and the method comprises the steps: receiving simulation parameters of a laser radar sensor, assembling the simulation parameters into a constant buffer area object, and carrying out the simulation of the laser radar sensor on the constant buffer area object; the method comprises the following steps: allocating a structured point cloud buffer area and binding shader resources in a rendering dependency graph framework, scheduling and calculating a shader to generate a two-dimensional thread index, calculating a light beam direction vector corresponding to each thread and constructing a ray descriptor based on the two-dimensional thread index, and obtaining hit information; the method comprises the steps of calculating echo intensity based on hit information according to a preset intensity model, simulating noise, writing point cloud attributes into a structured point cloud buffer area, and asynchronously looking back point cloud data for consumption through the structured point cloud buffer area. And the throughput and the data consistency are improved.

The present disclosure relates to a data rendering method for a real-time interactive three-dimensional Gaussian splatting scene. In the method, longitudinal and lateral block-wise partitioning is performed on a three-dimensional Gaussian splatting scene on the basis of a non-solidifying progressive construction technique, so as to optimize streaming media loading performance; furthermore, by means of scene block compression and a priority-based transmission and loading technique based on a user visual perception center, the data transmission efficiency and online real-time rendering efficiency of the three-dimensional Gaussian splatting scene can be effectively improved even under bandwidth-limited conditions, thereby satisfying real-time interaction requirements and improving the user experience even in network fluctuation environments.

The invention provides an electric power communication SDH network operation state real-time monitoring system and method, and the system comprises a multi-protocol adaptive collection module which carries out the collection of network element information, performance information, alarm information and equipment information of an electric power communication SDH network according to a preset strategy, performing standardization processing on the acquired multi-source heterogeneous data through a protocol analysis engine module to obtain structured data; the intelligent data screening and deconstruction module is used for analyzing, associating and identifying the structured data to obtain a key business entity; the lightweight packaging module is used for carrying out serialization and compression optimization on the key business entities to obtain lightweight packaging data; and the rendering module is used for carrying out real-time rendering on the basis of the lightweight packaging data to obtain a real-time visual operation state of the electric power communication SDH network. According to the invention, the acquisition efficiency is optimized, real-time monitoring is supported, the bandwidth cost is reduced, invalid data transmission is reduced by 80%, and the fault positioning time is compressed from an hour level to a minute level.