50 results about "3d localization" patented technology

The application discloses a real-time leakage positioning monitoring system and method based on electrical resistance tomography, relates to the technical field of process tomography and industrial safety monitoring, and solves the technical problem that it is difficult to intelligently identify, three-dimensionally position and quantitatively evaluate leakage events; the system comprises a sensor array, a multi-channel data acquisition and excitation module, a real-time data processing and image reconstruction module, a leakage intelligent identification and positioning module, a man-machine interaction and data management module; the core lies in that a multi-layer nested annular array and a sparse measurement strategy are adopted to improve data acquisition efficiency, a physical information neural network is introduced to reconstruct an image, a measurement topological coding matrix and a physical residual correction module are used to reconstruct the image, and finally, a time-space dual-domain leakage probability field model is constructed, a virtual sensor network and an improved D-S evidence reasoning framework are combined to intelligently identify and position a leakage diffusion mode.

The application relates to a Beidou RTK measurement point cloud self-adaptive filtering processing method, belonging to the field of high-precision three-dimensional positioning technology based on the Beidou system. The method firstly determines whether it is suitable to directly take the median filtering result as the filtering output according to the data amplitude span state of any measurement point in the Beidou RTK measurement point cloud and the effective satellite quantity when the point cloud is acquired, and quantifies the reliability of the median filtering result through the data fluctuation and direction change of the local neighborhood of the measurement point when it is not suitable, so that the fusion filtering result determination of the median filtering result and the original measurement value is completed in the case that the direct median filtering result is not suitable for filtering output. The application can avoid the reverse filtering optimization effect of the median filtering in the Beidou RTK measurement point cloud, and improve the RTK measurement point cloud data filtering accuracy of the median filtering.

This invention relates to the field of orchard navigation technology, specifically a 3D method for orchard operations based on a grid map. The method includes: constructing an orchard point cloud map using a mapping algorithm; filtering the acquired 3D point cloud map and constructing a triangular grid map of the orchard, embedding the orchard's GPS information into the triangular grid map; based on the triangular grid map with GPS information, performing ICP registration between the point cloud data scanned by the LiDAR and the points on the triangular grid map using the RMCL 3D positioning algorithm, outputting the 3D pose of the operation robot in the orchard; updating the 3D pose of the operation robot in the orchard using GNSS information; planning the navigation path of the operation robot using grid navigation based on the GPS information of the target point; and controlling the operation robot to dynamically avoid obstacles and navigate to the target point. This invention enables precise positioning and autonomous navigation of the operation robot in hilly orchards, allowing it to complete tasks such as orchard inspection, plant protection, harvesting, and transportation.

This invention discloses a method for autonomous operation of humanoid robots based on ROS. It employs a decision-making and control framework that fuses hierarchical finite state machines with YOLOv8 vision to achieve continuous autonomous execution of multiple tasks in complex unstructured environments. The method includes: system initialization and standardized processing of multimodal perception data; target 3D localization fused with YOLOv8 and coordinate transformation; task scheduling and state transitions using hierarchical finite state machines; segmented navigation and motion planning of joint trajectories; and ROS command issuance and full-process closed-loop control. This invention addresses the rigidity of traditional script-based control by fusing hierarchical state machines with visual servoing. Combined with a full-process anomaly handling mechanism and multi-target scene processing rules, it can dynamically adapt to changes in complex unstructured environments, effectively avoiding problems such as state freezes and erroneous grasping.

This invention relates to the field of engineering cost technology, and discloses a method and system for generating cost schemes driven by a component system. The method includes: extracting annotation information from 2D CAD drawings and identifying the spatial positioning attributes of components; generating 3D bounding boxes for component positioning based on multi-view geometric projection rules; inferring spatial adjacency relationships of components and calculating adjacency geometric parameters through bounding box intersection detection; expanding the component system to add spatial adjacency trigger dependency definitions; parsing bill of quantities items to match adjacency records and calculating linked quota quantities; and generating a cost combination scheme including linked quotas. This invention solves the technical problem of not being able to automatically identify and measure linked quotas dependent on spatial adjacency relationships when a BIM model is lacking.

A kind of open pit model dynamic updating method based on real-time positioning of excavator and SDF belongs to the technical field of digital mine. The initial pit triangular mesh model is voxelized to generate a three-dimensional symbolic distance field based on the definition of vertical coordinate symbol. The three-dimensional positioning data of the center of rotation of the excavator is collected in real time through the GNSS-RTK positioning module, and the working state is judged based on the preset time window and displacement threshold. When the excavator switches from displacement state to digging state, a spherical mining influence domain is constructed with the current positioning data as the center of the sphere and the maximum digging radius as the radius, and the voxel symbol distance field value in the domain is set to 0. A control point set is constructed by boundary constraint points and background constraint points, and global symbol distance field smoothing reconstruction is realized by using radial basis function interpolation. The zero value isosurface is extracted by moving cube algorithm to generate the updated pit model. The present application realizes high-frequency model updating synchronized with operation, has low deployment cost and strong robustness, and solves the problems of model updating lag and visualization penetration.

The application provides a deep foundation pit anti-floating risk assessment method based on monitoring data, and relates to the technical field of deep foundation pit anti-floating. The method comprises the following steps: carrying out layered multi-source heterogeneous monitoring on the deep foundation pit, obtaining a water pressure deformation data set and extracting spatial node and time sequence parameters; calculating the spatial distribution gradient of buoyancy and the dynamic resistance threshold; then establishing a local mechanical coordinate system in the effective stress coupling area, extracting a structure deformation mutation sequence from the data set; subsequently constructing a space-time coupling graph, generating a nonlinear attenuation path covering a weak area through a feature fusion algorithm; finally extracting a local anti-floating risk set along the path, generating a global anti-floating evolution graph through time domain smoothing and three-dimensional space mapping, and outputting an evaluation result containing three-dimensional positioning and automatic hardware control instructions. The application realizes space-time linkage analysis of foundation water pressure and structure deformation, and significantly improves the fine early warning and entity control ability of anti-floating risk under complex working conditions of deep foundation pit.

This application discloses a multimodal data collaborative intelligent detection and 3D localization method for maritime vessels, comprising: fine preprocessing of collected multi-source heterogeneous data and preliminary target detection; dynamic weighted fusion of the preprocessed multi-source heterogeneous data, combined with depth data from lidar, to achieve accurate coordinate recovery of the vessel target in 3D space; and prediction and updating of the motion state of the identified target based on a multi-tracking mechanism collaborative strategy. By fully utilizing the rich texture information of visible light images, the anti-interference capability of thermal imaging, and the precise spatial depth data of lidar through multimodal data perception and preliminary processing, heterogeneous data deep fusion and 3D localization, and advanced multi-target continuous tracking and state estimation, this method integrates a multi-tracking mechanism collaborative strategy to predict and update the motion state of the identified target, effectively addressing the problem of tracking interruption caused by nonlinear motion or temporary occlusion of maritime targets.

This invention provides a method, system, and medium for three-dimensional positioning and displacement measurement that integrates microwave and vision technologies. The method includes: rigidly connecting at least one microwave sensor and at least one vision sensor as a joint sensing unit; establishing a spatial transformation relationship between the coordinate systems of the two sensors through calibration; fusing the target's radial distance and angle information provided by the microwave sensor with the target's two-dimensional imaging information provided by the vision sensor; and directly solving for the target's three-dimensional coordinates in the selected coordinate system using an optimization algorithm; further combining the high-precision one-dimensional displacement information from the microwave sensor with the two-dimensional displacement information from the vision sensor to reconstruct the target's complete displacement vector in three-dimensional space. This invention, through a compact heterogeneous sensor fusion architecture, effectively solves the problems of limited dimensionality of single-sensor modes, high cost of multi-sensor networking, system complexity, and poor engineering applicability, achieving high-precision, low-cost, and full-dimensional non-contact three-dimensional positioning and displacement measurement.

This invention relates to the fields of artificial intelligence and computer image recognition technology, and discloses an intelligent method and system for detecting the quality of interior decoration. The method includes: applying a composite acoustic signal through a multi-band acoustic excitation device, and simultaneously acquiring multi-channel acoustic response, infrared thermal imaging sequences, and electromagnetic induction data; extracting acoustic impedance abrupt changes, thermal conduction anomalies, and electromagnetic continuity features respectively; generating multimodal embedding vectors using three convolutional neural network encoders, and generating a unified representation through a cross-modal attention fusion module; finally decoding and outputting defect type, confidence level, and three-dimensional positioning information, and superimposing it onto a building information model to generate a visual report. This system integrates a multi-physics sensing unit and an edge computing module to achieve non-invasive, fully automatic, and high-precision interior decoration quality detection. This invention improves defect detection rate and positioning accuracy through multimodal fusion and deep learning, increasing detection efficiency compared to manual methods.

This invention provides a method and apparatus for visual detection and harvesting control of guava maturity, relating to the field of visual detection technology. It acquires multi-view time-series images from a drone and extracts individual guava image blocks and depth information to determine spatial location. A feature encoder extracts feature vectors and performs clustering to determine maturity levels. Then, an optimization problem is constructed to plan the harvesting path, aiming to maximize harvesting revenue per unit time. Finally, the drone and robotic arm are controlled for flexible harvesting. This method can achieve automatic identification of guava maturity, 3D positioning, revenue-driven target selection and path planning, and precise flexible harvesting even in the absence of numerous manual maturity labels. It forms an integrated autonomous harvesting system with a closed loop of perception, understanding, decision-making, and execution, effectively solving problems such as difficult feature extraction, high manual labeling costs, low planning efficiency, and fruit damage in greenhouse guava harvesting.

This invention discloses a machine learning-based method, device, and medium for 3D model object recognition, relating to the field of 3D model recognition technology. The method includes: extracting topological features from a set of 3D ROI (Region of Interest) bodies to generate topological constraint codes; simultaneously performing shape decomposition to generate shape syntax codes; integrating the topological constraint codes and shape syntax codes to generate a manufacturing syntax dual-code description package; classifying the manufacturing syntax dual-code description package and interpreting it to generate 3D semantic segmentation; calculating the overlap degree by projecting onto a standard CAD view to generate a manufacturing conformity certificate; extracting the 3D semantic segmentation from the manufacturing conformity certificate; mapping the 3D semantic segmentation to the CAD 3D model to generate a direction annotation layer; associating this layer with the manufacturing conformity certificate to generate a manufacturing semantic object recognition list. This invention achieves structured encoding of manufacturing semantics, enhancing the interpretability and industrial applicability of 3D model recognition.

This invention discloses a method for real-time 3D tracking and localization of microscale targets in a bright-field microscopic environment, relating to the fields of image recognition and reconstruction technology. The method includes inputting continuous microscopic images into a target detection model, outputting the target's 2D position, regressing the defocus height and direction from a single frame of local images using a fine-grained ROI-based defocus regression model, directly using the model output to drive a 2D motion platform and a focusing device, fusing the 2D position information of each frame with the corresponding defocus axis information, and incorporating the position information of the previous frame as a priori using a temporal correlation mechanism to achieve continuous correlation of the microscale target in the time dimension, thereby stably reconstructing the target's 3D position and outputting a smooth real-time motion trajectory. This invention achieves real-time 3D localization and stable tracking of microscale targets under single-view bright-field microscopic conditions, ensuring both positioning accuracy and system real-time performance and accuracy.

This invention provides a vehicle 3D localization method, device, vehicle, and storage medium based on monocular vision. It acquires images in real-time from a monocular camera, processes them using a multi-task recognition model to obtain first target information, acquires parameter information corresponding to the monocular camera, and calculates second target information of the target vehicle based on the first target data. Finally, it tracks the target vehicle based on the second target information. Even in environments with limited texture, it can obtain camera poses with high accuracy, thereby achieving high-accuracy vehicle poses. This reduces the impact of environmental texture variations on vehicle localization, significantly improving the robustness of the vehicle localization system. By using the same scale feature layer for vehicle sideline detection and vehicle detection, each feature point corresponds to a set of detection results, eliminating the need for subsequent matching and improving target tracking efficiency.

This application discloses a monocular 3D localization method and apparatus that combines Phy-MSFF with physical geometric priors and a multi-scale feature fusion network, relating to the field of computer vision. The method includes: firstly, constructing a physically interpretable geometric depth baseline using keypoints; then, jointly inputting multi-scale appearance features and geometric features into a lightweight regression network to predict depth residuals; and finally, correcting the geometric depth baseline using these residuals. This fully utilizes the complementary information of pedestrian appearance and geometric priors to achieve precise 3D localization of the pedestrian target. Furthermore, a hierarchical degradation strategy is introduced to address long-tail problems such as pose loss.

The application provides a fruit and vegetable picking robot based on UWB positioning and YOLOv5 vision fusion and a control method thereof, and relates to the technical field of agricultural robots.The robot comprises a tracked or wheel-track combined mobile chassis, a six-degree-of-freedom mechanical arm and a flexible end effector for fruit picking, a UWB high-precision positioning module for global positioning, a binocular depth camera for target three-dimensional positioning, a laser radar obstacle avoidance module for environmental obstacle avoidance, an embedded cooperative control system, a communication module and a power supply module for power supply.The application adopts a dynamic fusion architecture of UWB global positioning and YOLOv5 visual local recognition.Compared with a traditional single sensing scheme, through the cooperative mode of global positioning bottom navigation and local visual accurate calibration, the problem of single visual positioning drift and the cumulative error of the odometer in long-distance driving is solved, and the accuracy of the robot orchard inter-row cruise, path correction and fixed-point parking is greatly improved.

The application relates to the technical field of computer vision, in particular to a three-dimensional positioning method and system of array elements, wherein the method comprises the following steps: acquiring a multi-view high-bit-depth gray image frame sequence of an array element to be measured; completing dynamic range conversion from high-bit-depth to low-bit-depth through adaptive bit-depth mapping enhancement; performing primary detection on the low-bit-depth image to obtain a candidate coded visual marker and a corresponding ROI region; performing local mapping enhancement on each ROI region to obtain an enhanced ROI image; performing secondary precision detection on the enhanced ROI image to output a target coded visual marker containing a target corner point and a single-frame pose; determining a corner point observation covariance according to the ROI dynamic range; constructing an incremental factor graph taking the single-frame pose as a state quantity and taking a re-projection error as an observation constraint, establishing an adaptive noise-weighted multi-frame joint optimization target; and performing optimization solving based on the optimization target to obtain three-dimensional coordinates of the array element in a world coordinate system. Through the application, the precision, consistency and real-time performance of three-dimensional positioning of the array element are improved.

This application relates to a navigation method and system for underground parking lots based on multi-source fusion. The method includes acquiring a 3D building information model of the underground parking lot and constructing a 3D navigation topology map based on the 3D building information model; collecting multi-source data, performing preprocessing operations on the multi-source data to obtain preliminary multi-source positioning results; evaluating the quality indicators of the preliminary multi-source positioning results to obtain multi-source positioning quality evaluation results, and performing dynamic weighted fusion calculations based on the quality evaluation results to obtain 3D positioning results representing the user's real-time location; and performing path planning based on the 3D positioning results and the user's target location, combined with the 3D navigation topology map, to obtain the target navigation path. This application improves the accuracy of navigation in underground parking lots.

The application discloses a tunneling target positioning method based on complex loose body multi-modal data survey, relates to the technical field of underground space artificial intelligence survey, and comprises the following steps: synchronously collecting multi-modal sensing data and tunneling working condition data in a tunneling process and performing adaptive noise reduction processing; extracting an energy envelope feature sequence of the multi-modal sensing data and performing depth mapping alignment; inputting the aligned multi-modal data features into a double-branch feature coding network, introducing a semantic prior vector as a context constraint, distributing fusion weights of the multi-modal data features through a cross attention mechanism, and outputting a high-dimensional fusion feature tensor; and constructing a physical information constraint neural network, taking electromagnetic and acoustic physical equations and auxiliary sensor data as constraint terms of a loss function for calculation, and outputting three-dimensional positioning coordinates of a tunneling target in a tunnel global coordinate system. The method eliminates time-depth conversion errors caused by complex loose body multi-variation.

The application discloses a geotechnical engineering monitoring system based on 3D printing prefabrication and a construction method thereof. The geotechnical engineering monitoring system comprises a grid base, a sensor and a data acquisition unit. The grid base is formed by 3D printing prefabrication. Sensor cabins are arranged at part of the intersecting nodes of the rib strips of the grid base. The sensor cabins are of a groove type structure. The sensor is arranged in the sensor cabin and connected with the data acquisition unit. The grid base is a stress contour gradually changing type topological grid. In addition to the rib strip intersecting nodes provided with the sensor cabins, the rest of the rib strip intersecting nodes are all provided with upward extending bionic root system structures. The grid base integrated with the sensor cabins has mechanical reinforcement and intelligent sensing functions, realizes high-precision three-dimensional positioning of the sensor, full-life-cycle physical protection and mechanical continuity interlocking function with the geotechnical body.

A system and method are provided for providing continuous automatic alignment of bone(-pieces), instruments, and implants based on only one additional X-ray image without the need to change the position of the X-ray source. According to the invention, a 2D X-ray image is received, which shows the target surgical site. Artificial intelligence-assisted automatic detection and localization of objects such as bone(-pieces), tools (such as drills), and / or implants, combined with prior information about the surgical procedure, is used to calculate their relative 3D positions and orientations at any desired time. This alignment of objects can be repeated multiple times, thus providing guidance in quasi-real time.

Disclosed is an editing system that accounts for or leverages the three-dimensional (“3D”) positioning of 3D image data to edit attributes of a selected first set of image data based on attributes of an unselected second set of image data that is determined to be a threshold distance from the first set of image data and on the same surface as the first set of image data. The system leverages x, y, and z coordinates as well as surface normals to exclude image data from the editing that is within the threshold distance but that forms part of a different object, surface, or side about an edge of a surface than the first set of image data. The system also modifies the attribute adjustment based on the distance separating the first set of image data from a render position or each instance of the second set of image data.

The application provides a flexible assembly test method under visual guidance, and focuses on the field of flexible assembly of intelligent manufacturing equipment. The core research and development of the application are 3D and 2D AI visual guidance robot flexible assembly test complete technology and equipment, which are fused with the cross application of AI visual identification, robot motion control and flexible assembly process technology, and include the following: an integrated scheme of a 3D and 2D AI visual guidance camera, an industrial robot and a flexible end effector, cooperative operation of 3D visual three-dimensional positioning and 2D visual accurate checking, and flexible assembly test suitable for multi-specification workpieces. The application aims to solve the problems of insufficient cooperative control precision of vision and robot, low fusion efficiency of multi-modal visual information, lack of flexible assembly test scheme suitable for multi-variety and variable batch production, and restriction of quality improvement and efficiency increase and intelligent upgrading of an assembly production line.

This invention relates to the field of ship perception, specifically to a data analysis-based intelligent monitoring system and method for multimodal ship perception, comprising: an image processing module, a sample calibration module, a modality fusion module, a target localization module, and an automatic formation module. The image processing module is used to acquire multimodal images; the sample calibration module is used to extract ROIs and obtain standard samples for each modality; the modality fusion module is used to construct a three-modal dataset; the target localization module is used to perform three-dimensional localization of targets; and the automatic formation module is used to control ships to avoid obstacles and form formations. This invention can improve the robustness and accuracy of shipborne obstacle detection, enhance the tracking performance of target tracking algorithms, improve the stability of obstacle target detection, improve the quality of high-resolution point clouds, enhance image fusion effects, improve the response speed of ship navigation, and improve the operational efficiency and safety of ship swarms.

A visual positioning method for an installation manipulator for a support arm of an overhead line network, characterized in that it comprises the following steps: Step 1: Capturing an image of an installation location using a depth camera and preprocessing the image; Step 2: Inputting the preprocessed image into an enhanced YOLOv8 network model for feature extraction and recognition; Step 3: Inputting the 2D image recognized in Step 2 into an image processing module, wherein the image processing module combines the depth information obtained from the depth camera to obtain 3D positioning data; Step 4: Transmitting the 3D positioning data obtained in Step 3 in real time to a manipulator control module, wherein the manipulator control module performs path planning to generate a motion trajectory and control instructions to direct the manipulator to execute a corresponding movement;Step 5: The image processing module continuously focuses on a target position during the movement of the manipulator, performs error detection, and adjusts the movement of the manipulator in real time; Step 6: The depth camera, the manipulator control module, and the image processing module send the on-site image and an installation situation back to a control terminal, and the operator monitors the on-site installation situation via the control terminal and intervenes manually via the control terminal;wherein the improved YOLOv8 network model comprises an input layer, a CSP multi-scale feature extraction module, an environment-adaptive feature extraction module, a multi-head self-attention mechanism, and a precise positioning optimization module, wherein the environment-adaptive feature extraction module is embedded within the CSP multi-scale feature extraction module, and wherein the image is output sequentially through the environment-adaptive feature extraction module, the multi-head self-attention mechanism, and the precise positioning optimization module;wherein an environment-aware unit of the environment-adaptive feature extraction module uses global average pooling and a full interconnect layer to extract global environmental information from an input feature map, and dynamically adapts a feature extraction strategy of the network in different environments, wherein the precise positioning optimization module improves the positioning accuracy of the support arm of the overhead line network through anchor point optimization, structure perception, and an improved positioning loss function; wherein a specific processing procedure of the environment-adaptive feature extraction module is as follows: after the input feature map is processed by the global average pooling and the full interconnect layer, the input feature map is transformed into an environment feature vector;Based on the environment feature vector, an original feature map is adaptively modulated using a specific formula as follows: e=sigmoid(W2⋅ReLU(W1⋅GAP(X)))X'=X□(1+γ⋅e) where W1 and W2 are a weight matrix of the full connection layer, where sigmoid is an activation function, where X is the input feature map, where e is the environment feature vector, where X' is a modulated feature map, and where γ is a learnable scaling factor, where □ denotes a channel-by-channel multiplication operation;wherein the feature maps modulated by the environmental feature vector are processed by dilation folding with different expansion rates, with a specific formula as follows: C=[Convd1(X'),Convd2(X'),Convd4(X')] where Convdx denotes the dilation folding operation with a 3*3 expansion rate of X, where d1, d2 and d4 each denote different expansion rates, where C represents a splicing result of the feature maps processed by dilation folding with three different expansion rates; wherein the multi-scale features are fused by 1*1 folding to produce a final feature map F.;

This invention discloses a fire and smoke recognition method that integrates temperature data and YOLO target detection, belonging to the field of intelligent environmental monitoring technology. This invention creatively constructs a multi-channel YOLO input by accurately registering and fusing visible light and infrared temperature channels. This not only effectively eliminates the interference of photothermal artifacts in complex backgrounds but also significantly improves the robustness of fire and smoke boundary detection. Furthermore, by introducing a dark channel prior step in the detection box processing stage, it overcomes the limitation of traditional visual networks that can only perform qualitative box selection, achieving real-time calculation of dynamic physical concentration of smoke. This method deeply correlates the multi-dimensional fire attributes calculated at the front end with the back-end UAV, directly outputting 3D geographic positioning through coordinate transformation. It achieves full-chain automation from qualitative perception and quantitative measurement to spatial mapping, providing technical support with significant practical value for smart environmental protection.

The present invention relates to a computer-implemented method for 3D localization of an object based on image data and depth data, the method applying a convolutional neural network having a first set of successive layers, a second set of successive layers and a third set of successive layers, each layer being configured to have one or more than two filters, the convolutional neural network being trained to associate recognition / classification of an object and 3D localization data of the recognized / classified object with an image data item and a depth data item, the method comprising: applying the first set of layers to extract one or more than two image data features from the image data; applying the second set of layers to extract one or more than two depth data features from the depth data; fusing the one or more than two image data features and the one or more than two depth data features to obtain at least one fused feature map; applying the third set of layers to process the at least one fused feature map to recognize / classify the object and to provide 3D localization data of the recognized / classified object.

The application relates to a near-distance positioning and long-distance detection transformation method and system based on camera and triangle transformation fusion, and belongs to the technical field of computer vision and visual positioning. In the long-distance stage, stable detection is completed by relying on outer large-scale markers; in the detection transformation stage, a detection algorithm self-adaptive transformation is triggered based on pixel scale and a geometric model; and in the near-distance positioning stage, high-precision three-dimensional positioning is completed based on inner small-scale markers and a triangle transformation model. The application introduces a detection transformation mechanism facing distance changes, preferentially utilizes outer visual markers to complete stable detection in the long-distance stage, and then switches to inner visual markers for accurate positioning after the target gradually approaches, so that the effective detection distance range of the system is significantly expanded, and the continuity and stability of the long-distance and near-distance switching process are improved without reducing the near-distance positioning accuracy.