788 results about "Multiple view" patented technology

This invention provides a virtual reality method and system for constructing a geometrically perceptible material field based on a large model and driving real-time interactive sound synthesis. The method includes: acquiring geometric data and multi-view observation data of the target object; generating prior material constraint information from a large model and fusing the multi-view data to obtain the material feature field of the target object; generating physical sound assets offline based on the material feature field; acquiring interaction state information during virtual reality scene runtime, mapping it to excitations consistent with geometric position and interaction process, and synthesizing collision and friction sounds, enabling the same object to produce continuously changing and distinguishable sound feedback under different touch positions, sliding paths, and contact areas. This invention achieves real-time synthesis and output of collision and friction sounds by constructing a material feature field bound to the geometric space of the virtual object and driving physical sound assets in conjunction with interaction state information during runtime, thus enhancing the consistency of auditory feedback and immersive experience in virtual reality scenes.

The application relates to the technical field of intelligent auxiliary reminding, in particular to an auxiliary reminding method based on a smart wearable device, which comprises the following steps: collecting behavior data of the elderly through a multi-view camera, an infrared sensor and a pressure sensing device, constructing a four-dimensional characteristic behavior sequence graph, and using a deep learning model with a multi-modal fusion and attention enhancement mechanism to analyze behavior semantics and generate graded reminding instructions. The application can accurately monitor the daily behavior of the elderly, dynamically adjust the reminding strategy, trigger an emergency alarm in an abnormal situation and record intervention logs, and effectively improve the life safety of the elderly and the monitoring efficiency.

Provided are a multi-dimensional fish school perception method and system, and an electronic device. The multi-dimensional fish school perception method includes: acquiring a multi-view optical image and original sonar data of a fish school by using the multi-dimensional fish school perception apparatus; obtaining a first 3D (three-dimensional) feature map of the fish school based on the multi-view optical image; obtaining a second 3D feature map of the fish school based on the original sonar data; extracting fused features based on the first 3D feature map and the second 3D feature map to obtain a fused feature map; and obtaining, based on the fused feature map, a 3D detection result of a target fish school by using a self-attention-mechanism-based 3D target detection method.

The application discloses a pose-deformation joint estimation method and system for micro contact assembly process, a workpiece is scanned by a mobile device from multiple perspectives, and a global point cloud model of the workpiece is reconstructed. Next, the overall pose parameters and the local surface topography of the workpiece are solved synchronously, and then a displacement field and a strain field are constructed to quantitatively represent the deformation state. Subsequently, the image, point cloud, pose and deformation parameters are fused, and a pre-trained joint estimation model is used to realize rapid prediction of the current coupling state of the workpiece. Through the built-in consistency verification and online updating mechanism, the reliability of the prediction result is ensured, and the final reliable state estimation is converted into a robot assembly correction instruction in real time, thereby forming a closed loop of perception, estimation and control. The application can realize sub-millimeter level pose measurement and high-precision deformation prediction, significantly improve the dynamic adaptability, robustness and automation level of the assembly process, and is suitable for automatic assembly of high-precision components in the fields of aviation, aerospace and the like.

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.

The application relates to a kind of city building agent reconstruction methods based on aerial image.The method comprises: obtaining the multi-view image corresponding to target building;Multi-view image includes near vertical view image and inclined view image;According to near vertical view image, the bottom contour of target building is determined;Determine the first depth difference between each pixel point in inclined view image;According to the camera pose parameter of multi-view image and the current predicted height of target building, rendering is carried out, and the rendering depth map of target building under the current predicted height is obtained;With the distance between the second depth difference of each pixel point in the rendering depth map and the first depth difference as the optimization target, the current predicted height is optimized, and according to the bottom contour and target measurement height, the three-dimensional building model corresponding to target building is constructed.Using the method can improve the efficiency and accuracy of city building agent reconstruction.

The application discloses a multi-view human body reconstruction method based on photometric consistency matching and optimization algorithm, which comprises the following steps: firstly, obtaining a rough human body surface through a visual hull algorithm according to a human body mask image; secondly, using photometric consistency constraint to optimize the shape initialized from the visual hull, so as to obtain a dense human body surface model; thirdly, calculating an illumination coefficient by using the diffuse reflection principle; and finally, using a light and shade optimization algorithm to perform high-speed real-time rendering on the dense human body surface model, so as to obtain a final simulation human body model. The application can optimize the initialized rough surface by using the contrast of the gray scale image, maintaining the photometric consistency constraint and the differentiable rendering, and can effectively solve the problems of the unsmooth surface, the unobvious geometric details and the color estimation by using the diffuse reflection principle to estimate the diffuse reflectivity and the illumination. The application can optimize the human body surface by the difference of the light and shade and the color in the image.

This invention relates to the field of shield tunnel segment inspection technology, and discloses a visual inspection-based shield tunnel segment assembly machine assembly quality inspection device and method, including: achieving autonomous perception and rapid recovery of calibration parameters when they fail due to vibration or temperature drift through time-efficiency monitoring and sliding window optimization technology; overcoming calibration bottlenecks through texture perception and active exploration trajectory planning technology; solving the calibration instability problem caused by the lack of natural textures by artificially creating multi-view geometric constraints using micro-motion trajectories; achieving intelligent decision-making and adaptive adjustment in the active exploration process through information gain-driven convergence judgment and global optimization technology; significantly enhancing parameter estimation confidence through global optimization integrating multi-source data; and constructing an unattended autonomous calibration management system through result encapsulation and closed-loop feedback technology, enabling persistent storage of calibration parameters, accumulation of prior knowledge, and real-time timeliness monitoring.

A combined convolutional neural network driven automobile flow field excitation prediction method comprises the following steps: 1) constructing a plurality of automobile deformation models with continuous geometric feature changes, and sampling multi-view images of all automobile deformation models to construct an image dataset; 2) simulating transient flow fields of all automobile deformation models, obtaining corresponding fluid dynamic pressure excitation and sound pressure excitation, and labeling corresponding multi-view images in the image dataset with excitation labels to construct a training set; 3) constructing a combined convolutional neural network model; 4) taking multi-view images as input and excitation labels as output, and training the combined convolutional neural network model by using the training set; and 5) obtaining multi-view images of a to-be-predicted automobile, inputting the multi-view images of the to-be-predicted automobile into the trained combined convolutional neural network model, and obtaining predicted fluid dynamic pressure excitation and sound pressure excitation. The application can significantly reduce the dependence of automobile flow field excitation calculation on supercomputing services and reduce wind noise development costs.

The present application relates to the field of three-dimensional reconstruction, and provides a pig weight estimation method, device, equipment, medium and computer program product.The method comprises the following steps: acquiring multi-view image data of a pig to be detected; generating sparse point cloud data based on the multi-view image data; generating a multi-view rendering image based on the sparse point cloud data and three-dimensional Gaussian data; the three-dimensional Gaussian data is converted from the sparse point cloud data; inputting the three-dimensional Gaussian data into a weight estimation model to obtain pig feature information, and estimating the weight of the pig to be detected based on the pig feature information.The present application improves the accuracy of pig non-contact weight estimation and reduces the implementation cost through individual weight posture correction, weight scene three-dimensional reconstruction and the associated weight estimation method of fusing body posture semantics.

The invention discloses a virtual multi-view fusion millimeter wave point cloud imaging method, which belongs to the field of millimeter wave radars, and realizes three-dimensional point cloud imaging of a target by constructing a virtual observation view angle and fusing perception data of different view angles: establishing a multi-view synthetic aperture radar imaging model; a single-bit compressed sensing framework is introduced to construct a sparse imaging optimization problem, an accurate relay angle is solved by minimizing a model error, and a relay plane model error is corrected; and high-precision 3D point cloud imaging of the region of interest is realized based on the accurate angle. According to the method, a three-dimensional point cloud of a target is constructed by fusing virtual visual angle detection shielding characteristics formed by reflection of a plurality of relay surfaces, joint estimation of relay surface angle prior and 3D point cloud is realized by a nested iteration method, and artifacts and offset caused by prior errors are eliminated; the sparsity and the target continuity of the three-dimensional point cloud can be improved through the additional composite norm constraint, and the accuracy of virtual multi-view 3D point cloud imaging is effectively improved.

This invention provides an intelligent detection and classification method for oil and gas pipeline defects based on pseudo-color images, belonging to the field of pipeline safety inspection. First, multi-dimensional magnetic flux leakage data of the oil and gas pipeline in the axial, circumferential, and radial directions are obtained through a tensile experiment. The detection data is converted into visualized pseudo-color images, and image defect annotation is completed to construct a standardized pipeline defect dataset. In the detection stage, the dataset is input into a pre-trained YOLOv11 model to achieve accurate identification and location of potential pipeline defect areas. In the multi-view fusion stage, the defect bounding box output by the model is used to extract the target area of ​​the original pseudo-color image. Images of the same defect from multiple directions are stitched together to generate a multi-view pseudo-color image of the defect. In the classification stage, the Patch preprocessing module is used to complete image resizing and filling, and then the image is fed into a finely tuned DINOv3 classification model to efficiently complete the intelligent identification and classification of various pipeline defects.

The application discloses a multi-core cache consistency debugging method based on multi-view access and belongs to the field of multi-core processor debugging. The method is characterized in that a target address is specified in a main memory by a host computer through a JTAG debugging interface, five groups of data of CPU view, L1 data cache view, L1 instruction cache view, L2 view and real memory view corresponding to the address are read in turn, whether cache is invalid is judged by comparing data consistency, and a cache level inconsistent is marked. The debugging system comprises a built-in JTAG TAP controller, an enhanced debugging access module, a storage level access arbitrator and a data and state collector. The application realizes direct access of the JTAG interface to data of each level of cache, does not need to modify hardware or disable cache, does not interfere with system operation, can quickly locate cache consistency problems under a multi-core architecture, improves debugging efficiency and accuracy, and is suitable for debugging scenes of large-scale multi-core processors.

The application discloses a kind of variable working condition diagnosis methods of rotary reducer based on multi-view transfer, belong to fault diagnosis technical field.The method is first to the multiple visual vibration data collected is preprocessed and constructs variable working condition data set;Through fourier transform, extract each visual spectrum feature;Typical correlation analysis is used to construct embedding class discriminant transferable feature objective function, combined with maximum mean difference technique reduces the distribution difference between training and testing field, and introduces multi-view consistency constraint;Through generalized feature decomposition, solve common subspace projection, extract multi-view features with discriminant and transferability;Finally, nearest distance classifier is used to realize the fault diagnosis under variable working condition.The application effectively solves the problem that rotary reducer has poor generalization ability due to data distribution difference under variable working condition, improves the accuracy and reliability of fault diagnosis.

The application discloses a kind of training retrieval model method, retrieval method and device, it is related to retrieval and artificial intelligence technical field.The specific embodiment of the method includes: according to sample item information, generate first query text and second query text;Utilize the first query text training retrieval model includes question encoder;Utilize the second query text and the splicing text that the sample item information splicing of training document encoder included in the retrieval model;And utilize the retrieval model of training and carry out item information retrieval, the query text of different encoder in the training retrieval model is generated based on sample item information in the application embodiment, improve the richness and multi-view degree of training sample, improve the training effect of retrieval model, and improve the retrieval experience of user.

This invention discloses a method and system for large-scale federated 3D Gaussian sputtering scene reconstruction for resource-constrained edge networks. The method includes: constructing a resource-aware federated 3D-GS system consisting of a single edge cloud server and a set of ubiquitous edge sensing devices; on the edge sensing device end-to-edge side, based on the collected local multi-view real-scene image dataset and an adaptive lightweight mechanism, under given GPU memory budget and maximum latency constraints, adaptively selecting the optimal Gaussian point subset for local training, obtaining updated local spatial features, position gradients, and binary masks, and transmitting them back to the single edge cloud server via a wireless network; on the single edge cloud server end, using the local spatial features, position gradients, and binary masks uploaded by the device, mapping the structurally heterogeneous local model back to the global index space, performing structural consistency restoration, and then performing weighted aggregation to generate a 3D real-scene map.

This invention relates to the field of image processing technology and discloses a method for target 3D reconstruction and automated model generation. The method includes: acquiring multi-view original point cloud data of the target object, and obtaining a simplified point cloud dataset through voxel filtering and downsampling; dividing the data into smooth regions and sharp edge feature sets based on curvature characteristics, performing smooth fitting and edge-preserving reconstruction respectively, and then generating a composite mesh model through vertex merging and mesh stitching; performing topological structure analysis on the model, detecting and optimizing non-manifold elements and topological holes to obtain an initial mesh model; parametrically mapping and automatically unfolding the texture of the model surface geometric manifold characteristics, and fusing and binding the unfolded texture map with the initial mesh model to finally complete the 3D reconstruction of the target object; this invention can improve the efficiency of target 3D reconstruction and automated model generation.

This invention discloses a dual-alignment-driven zero-sample sketch-3D model retrieval method, belonging to the fields of computer vision and artificial intelligence. Addressing the problems of excessive modal gap and difficult semantic transfer between sketches and 3D models in zero-sample sketch-3D model retrieval tasks, this invention utilizes generative adversarial networks (GANs) to achieve cross-modal mapping from sketches to pseudo-views. A generator produces pseudo-views with structural consistency, narrowing the modal gap between multiple views of the sketch and the 3D model. Based on a dual-extractor structure, features are encoded for the sketch, pseudo-view, and multiple views of the 3D model, and normalization is applied to obtain high-dimensional embedding representations at a unified scale. A cross-batch memory mechanism is introduced to establish a FIFO queue to store historical features, and a cross-batch sample comparison loss is used to construct global semantic alignment, thereby expanding the range of negative samples and enhancing feature discriminative power. Simultaneously, a class center constraint is introduced through normalized Softmax loss to achieve global optimization of intra-class aggregation and inter-class separation.

This invention provides a method and apparatus for 3D tooth reconstruction based on multi-view geometric priors and specular Gaussian representation. The method first acquires multi-view image data containing teeth and gingiva. A multi-view geometric estimation module predicts and globally aligns dense 3D point maps to obtain an initial point cloud, camera intrinsic and extrinsic parameters, and a depth map. Based on this, a Gaussian primitive set is initialized and input into a specular Gaussian reconstruction module, where relevant parameters are combined to optimize the reconstruction. This module introduces depth and normal losses based on the color target, supervises and reduces the weight of highlight regions within the region of interest, and finally outputs the 3D reconstruction result. This invention can improve the stability and geometric accuracy of tooth reconstruction under high reflectivity and weak texture, enhance specular reflection and highlight expression capabilities, and balance the detailed reconstruction of tooth hard tissue with the geometric and appearance effects of gingival soft tissue, thereby improving the integrity of oral scene reconstruction and its clinical application value.

The application provides a plate tail intelligent grading equipment and method based on visual guidance, and the core is that through integration of multi-view visual perception, adaptive grabbing, digital twin and cloud-edge collaborative control technology, full-process intelligent management of plate tail from warehousing, grading storage to on-demand use is realized; multi-view images of the tail are collected by a camera array, three-dimensional point clouds are reconstructed through a stereo matching algorithm, and morphological characteristic parameters of the tail are extracted; with the help of digital twin technology, a virtual model of each tail is generated and synchronized, a stacking strategy is optimized based on a genetic algorithm, and space utilization is maximized; when a use request is received, the system can perform virtual cutting pre-performance in the digital twin environment, accurately match the demand, and use the stacked tail through the temporary storage-backfilling mechanism; finally, through the cloud-edge collaborative platform, visual monitoring and global scheduling optimization of the inventory and equipment state are realized, and the tail management efficiency and material utilization are significantly improved.

This application provides an action determination method and apparatus, a robot system, a storage medium, and an electronic device, relating to the field of robotics. The method includes: acquiring task instructions and multi-view image data, wherein the multi-view image data includes image data collected by the robot from M views, where M is an integer greater than or equal to 2; determining a feature set based on the multi-view image data, wherein the feature set includes a semantic feature subset, a geometric feature subset, and a three-dimensional spatial feature subset corresponding to the M views; and determining the robot's execution action based on the feature set and the task instructions. This solves the problem of inaccurate robot execution actions in related technologies, enabling action generation based on accurate three-dimensional spatial understanding and improving the accuracy of execution actions.

This invention relates to the field of image segmentation technology and discloses a construction site risk behavior recognition system based on big data. The system constructs a multi-view spatiotemporal representation of construction by uniformly aligning multi-view video streams, personnel positioning information, equipment operating status, and construction space area data at the construction site. Simultaneously, it constructs a segmentation model including an image encoder and a Bayesian mask decoder, and introduces a dynamic evidence adaptation mechanism based on Set Transformer and Hypernetwork to achieve dynamic segmentation of construction targets and hazardous areas in complex construction scenarios. Furthermore, by combining spatial posterior distribution, semantic posterior distribution, and a visual language model, it achieves reliability analysis and semantic discrimination of risk behaviors, thereby improving the reliability and adaptability of risk behavior recognition in complex construction scenarios.

The application relates to the technical field of surface defect intelligent detection, and discloses a surface defect detection method and device based on multi-view synthetic imaging. The method comprises the following steps: acquiring a multi-view synthetic image without dead angles of a detected target after quality evaluation; performing highlight suppression and shadow enhancement on the multi-view synthetic image to obtain an enhanced multi-view synthetic image; separating the enhanced multi-view synthetic image into corresponding multiple single-view images, respectively performing defect detection on each single-view image to obtain corresponding single-view defect detection results, and generating a final overall detection result based on the single-view defect detection results. Through multi-view synthetic imaging, image quality control, high-reflective surface enhancement processing and multi-view joint defect detection, the application can effectively solve the problems of light interference and defect missing detection in high-reflective surface detection, and can realize high-precision surface defect detection with no dead angle coverage in a single detection.

The application discloses a kind of multi-camera-based animal vital sign monitoring method, medium, equipment, it is related to animal vital sign monitoring technical field, method includes: using DLC posture estimation network to train animal multi-view video, obtain key point information;According to the key point information of main visual angle, the centroid motion of animal is analyzed and compared with motion threshold, determine animal state, in the continuous image sequence of relative static state time period, according to key point information, obtain region of interest;Corner point is detected in region of interest, and corresponding motion signal is acquired by tracking corner point;Correlation analysis is carried out to motion signal, and sorting is carried out based on information gain, and vital sign signal is obtained by fusing motion signal higher than set threshold;Power spectrum is used to analyze vital sign signal, obtain the energy distribution of vital sign signal, and the value of respiratory rate and heart rate is calculated.The application can effectively reduce artificial intervention, and has the role of promoting to animal experiment method innovation and data analysis.

This invention discloses an automatic mural splicing control system and method based on image recognition, belonging to the field of image recognition technology. The method includes: performing image recognition decomposition and structured encoding on the mural design drawing to obtain an assembly code map, a target wall reference rendering, and a 3D preview model; acquiring multi-view images of the wall work area and completing camera calibration, determining the transformation relationship between the camera coordinate system and the wall coordinate system to form a wall mapping matrix, and using the wall mapping matrix to map the assembly code map into a target placement coordinate table; acquiring pixel block images of the material area, extracting the appearance contour and appearance feature descriptors, and performing nearest neighbor matching with the assembly code map to obtain the corresponding entries in the target placement coordinate table. By combining the assembly code map with the target placement coordinate table, this invention can accurately generate the grasping and placement posture sequence of the robotic arm, achieving automated assembly and significantly improving the accuracy and efficiency of the assembly process.

The application discloses a rain weather multi-view feedforward Gaussian reconstruction method and system based on weather factor gating, and the method comprises the following steps: acquiring rain weather context images under different views and camera parameters thereof; inputting the rain weather context images into a weather factor decomposition network to generate predicted weather factors; constructing a matching reliability graph for representing the corresponding reliability degree across views according to the predicted weather factors; extracting multi-scale features of the rain weather context images; generating gated features according to the multi-scale features, the matching reliability graph and reliability gating; performing multi-view aggregation based on the gated features to generate fusion representation; inputting the fusion representation into a three-dimensional reconstruction network for feedforward prediction to obtain a three-dimensional Gaussian set; inputting the three-dimensional Gaussian set into a differentiable renderer to render a clean predicted image of a target view in combination with target view camera parameters; acquiring clean context images of multiple views to construct rain weather training samples; and constructing a loss function to train the weather factor decomposition network and the three-dimensional reconstruction network.