77 results about "3d coordinates" patented technology

This application provides a 3D reconstruction method, apparatus, device, and readable storage medium. An electronic device acquires multiple first images, each containing the background of a target scene and dynamic objects within that scene, with the sum of these first images capable of displaying a panoramic view of the target scene. The electronic device then performs masking processing on the regions containing dynamic objects in each of the first images to obtain an image set. Based on this image set, the background of the target scene is reconstructed to obtain a background point cloud, which is then used to generate a 3D reconstruction result of the target scene. This approach, by masking the regions containing dynamic objects in the first images of the target scene, eliminates the influence of feature point extraction and matching results on 3D coordinate estimation and camera extrinsic parameter estimation, ensuring successful reconstruction of the background point cloud of the target scene. This improves the success rate of 3D reconstruction and enhances the immersive conferencing experience.

The present teaching relates to obtaining three-dimensional (3D) information in a contactless manner using a one-dimensional (1D) distance sensor. The 1D distance sensor is associated with an origin from where a laser beam emitted in an orientation, where the origin / orientation are calibrated with respect to a marker attached thereon and tracked. In a surgery, the calibrated 1D distance sensor is tracked by a camera via the marker attached thereon. When an emitted laser beam hits at a location on a patient, the 3D coordinate thereof in the tracker coordinate system is determined based on tracked marker, the origin / orientation of the 1D distance sensor, and a distance between the origin of the 1D distance sensor and the location.

A system includes a projector that projects a light pattern having a time-domain modulation onto an object, the time-domain modulation including an integer number of adjacent modulation cycles. The system also includes a first time of flight (ToF) camera having a lens and a two-dimensional (2D) pixel array that converts light reflected by the object into an electrical signal. The at least one processor determines three-dimensional (3D) coordinates of a point of the object based at least in part on the electrical signal, the baseline length, the time-domain modulation, and the velocity of light in the air.

This invention proposes a method and system for detecting missing weld points on vehicle bodies based on 3D vision and global matching, belonging to the field of image recognition technology. The method includes: determining the transformation relationship between the camera coordinate system and the vehicle body coordinate system through calibration; at each observation position, a 3D camera acquires an image I containing a frame of 2D texture and the corresponding 3D data; inputting the acquired image I into a pre-trained deep learning object detection model, outputting the bounding box information (u, v, w, h) for each weld point; for each detected weld point, calculating the set S of the weld point's 3D coordinates in the vehicle body coordinate system. det For the theoretical solder joint library S cad Each theoretical solder joint P in cad,i In set S det The invention employs a "detect first, match later" strategy, which does not rely on the precise projection of the theoretical position, thus reducing the requirements for the positioning accuracy of the production line tooling fixtures and conveying system.

The present invention provides a system and method for a two-step registration procedure that enables complete surgical intervention by a robot. [Solution] A system for intraoperative navigation is provided, comprising a processing unit and a tracking device for real-time or near-real-time tracking of an object in a 3D coordinate system, wherein the processing unit is configured to receive an X-ray image depicting an anatomical structure including at least one bone, and to determine an incomplete registration of the position of the anatomical structure in the 3D coordinate system. The incomplete registration is determined based on a 3D dataset of the received anatomical structure and a plurality of virtual X-ray images generated from different directions based on the 3D dataset of the anatomical structure. The incomplete registration can be understood as an approximation of the position of the anatomical structure. Based on the incomplete registration, instructions for the navigation of the device are provided.

This invention discloses a human-machine collaborative safety monitoring method that integrates temporal features and multi-task risk assessment, belonging to the field of industrial human-machine collaborative safety protection. The method first constructs a static 3D point cloud template of the machine tool offline as a danger boundary, simultaneously acquires RGB-D images and performs pixel-level registration and fusion, and extracts initial features using a deep backbone network. It then uses a ConvLSTM convolutional long short-term memory network to model spatiotemporal features, suppresses instantaneous occlusion, and regresses to obtain a smooth and continuous 3D coordinate sequence of hand joints. Based on the coordinates, kinematic vectors are calculated and a hand gesture code of the work intention is generated. The multi-dimensional features are input into a multi-gate hybrid expert network (MMoE), and the weights are dynamically adjusted to output a comprehensive risk score. The system executes normal operation, linear deceleration, and hardware emergency braking according to the score, linking the machine tool PLC or robot controller. This enables high-precision hand positioning and occlusion tracking, predicts movement trends to reduce lag, balances safety and production efficiency, has strong industrial applicability, and is suitable for complex human-machine collaborative safety protection.

An automatic detection method for CBCT cephalometric landmarks based on multi-geometric guidance and specific perceptual coding includes the following steps: Step S1, downsampling the 3D CBCT image and obtaining a preliminary coordinate set through a coarse localization network; Step S2, cropping image blocks centered on the coordinates and extracting local features through a visual encoder containing a shared basic encoder and a low-rank adapter; Step S3, calculating the relative position matrix of the landmarks, encoding spatial relationships using radial basis functions, and constructing a multi-anatomical heterogeneous map; Step S4, inputting visual and edge features into a multi-geometric guidance Transformer, fusing global constraints and updating features using an attention mechanism; Step S5, extracting directional geometric relationships using spherical harmonic functions to construct higher-order update terms, and dynamically updating the heterogeneous map using a gated residual mechanism; Step S6, predicting coordinate offsets through a multi-layer network and performing iterative optimization to output high-precision 3D coordinates. This method significantly improves detection accuracy and robustness.

This invention discloses a robust visual servo control method for robots based on an adaptive update law, comprising: associating the pixel coordinates of feature points with their 3D coordinates in the camera coordinate system according to a pinhole camera projection model, and establishing a global linear regression model containing unknown position parameter vectors; designing an adaptive update law for the estimated values ​​of the unknown position parameter vectors based on the global linear regression model, used to estimate the unknown depth information parameters of feature points in real time, and reconstructing the visual Jacobian matrix online based on the estimated depth information parameters; constructing a visual state observer to obtain smoothed image feature estimates; defining a joint space-assisted tracking error based on the smoothed image feature estimates based on the reconstructed visual Jacobian matrix; and designing an integral sliding mode robust feedback term based on the integral sliding mode in the robot joint driving torque expression based on the lumped disturbance in robot dynamics, combined with feedforward compensation, to generate the final joint driving torque to drive the robot to perform servo tracking tasks. This invention can achieve highly robust and high-precision asymptotic tracking of target feature points under complex working conditions.

This invention relates to the field of 3D scanning, and more particularly to a 3D rotational scanning method, scanning device, and storage medium. By rotating a rotating platform by a first angle, a scanning device is mounted on the rotating platform. The rotating platform drives the scanning device to rotate, and the scanning device emits a laser. Based on the first angle and preset first rotation center coordinates, the laser's output port coordinates are determined. The first rotation center coordinates are the fitted center of multiple output port coordinates corresponding to various rotation angles of the rotating platform. The distance between sampling points obtained by laser scanning the sample surface is acquired. Based on the sampling point distances and the output port coordinates, the 3D coordinates of the sampling points in the world coordinate system are determined. A point cloud image of the sample is obtained based on the 3D coordinates of each sampling point on the laser-scanned sample surface. This method enables rotational scanning, correction and reconstruction of point cloud data, optimizes workspace utilization, and ensures high-precision measurement results.

A surgical system includes an instrument, tool, position markers, image sensor, and control unit. The instrument is configured to engage anatomical tissue along a trajectory that intersects an anatomical target location. The instrument's distal end engages the tissue. The tool manipulates instrument movements along the trajectory. The markers are disposed on a base surface on the tool. The image sensor is tool-mounted; defines a view field encompassing the base surface, markers, and target location; and transmits a live image stream. The control unit receives the live stream and has a processor that executes machine-readable instructions to identify visual spacing and orientation parameters of the markers in the live stream. The processor also calculates and registers in memory a reference location at the instrument's distal end in a 3D coordinate system and calculates distance offsets between the target and reference locations in the coordinate system and communicates the offsets to a user.

This invention discloses a method for 3D reconstruction of coaxial digital holographic particle fields combining structural tensors and guided filtering. Specifically, the method involves: normalizing the coaxial digital hologram and performing multi-distance numerical reconstruction to obtain a 3D reconstruction intensity stack; constructing a focus response using structural tensors and generating spatially consistent fusion weights through guided filtering, fusing them to obtain an extended depth-of-field image and a particle soft depth map; constructing a candidate score map based on the extended depth-of-field image to achieve candidate localization, adhesion separation, and diffraction ring suppression; and finely focusing and segmenting the measured particle size within the depth neighborhood, ultimately outputting the particle's 3D coordinates and size parameters. This invention aims to utilize the effective characterization capability of structural tensors for focusing features and the edge-preserving properties of guided filtering to construct a fusion method for multi-depth particle field reconstruction results. This method enhances particle target features while suppressing artifacts and background interference, improving particle localization and measurement accuracy, and achieving higher-precision 3D reconstruction of coaxial digital holographic particle fields.

The present disclosure relates to methods (200; 300; 400) for use in a moving platform for initialization of a visual-inertial odometry system (100) comprising a camera (1) and an inertial measurement unit (2), IMU. The methods comprise determining (201) point correspondence between at least a first image captured by the camera and a textured 3D model of an environment containing 3D coordinates given in a predetermined global coordinate system. The methods further comprise determining a relative rotation between the predetermined global coordinate system used for the camera and a IMU coordinate reference system, or estimating (210) a relative translation and rotation between the first and second images in a scale of the predetermined global coordinate system using the determined first and second point correspondences, said estimate indicating the scale for movement of the camera (1) for use in initialization of the visual odometry system.

Provided are a method for generating steerable realistic image content according to an embodiment and a motion simulation system thereof. A device for generating steerable realistic image content according to an embodiment includes at least one processor and a memory configured to store at least one program which is executed by the at least one processor, wherein the processor is configured to acquire a motion path in a realistic image captured by a camera while the camera is moving along a predetermined path, generate a 3-dimensional (3D) motion model based on the motion path, and synchronize the 3D motion model with the realistic image, and the 3D motion model includes a 3D moving object which moves along a 3D motion path generated along 3D coordinates of the motion path.

A method and a system are provided for determining a location of a brain target for neuromodulation therapy without requiring MRI imaging of the head. The system includes a sensor or a plurality of sensors configured to acquire data describing a geometry of a head of a subject. A processor is coupled to the sensor and is configured to receive the data describing a geometry of the head of the subject gathered by the sensor, determine 3D coordinates of head-surface contour points from the acquired data, apply a predictive model to the determined 3D coordinates of head-surface contour points, receive 3D coordinates of a located brain target as output from the model, and output 3D coordinates of the located brain target for subsequent delivery of neuromodulatory brain therapy. The model is pre-trained on a dataset of head and brain images and configured to associate head-surface coordinates with 3D coordinates of at least one deep brain target.

This invention discloses a track smoothness adjustment method based on 3D coordinates and multi-wavelength technology, aiming to solve the problems of accumulated adjustment errors caused by reliance on approximate models in existing technologies, and low efficiency due to the inability to simultaneously handle long and short wave irregularities caused by the single wavelength. The method includes: acquiring the designed and measured three-dimensional coordinates of the track; directly calculating the lateral and vertical versine deviations for each wavelength based on the three-dimensional coordinates using the moving string method according to a preset wavelength list; directly correcting the three-dimensional coordinates using a predetermined proportion of these deviations as a correction amount; completing a single-round adjustment by traversing the wavelength list, and performing multiple rounds of iterative optimization based on the updated coordinates to achieve multi-wavelength gradient smoothness of the track's three-dimensional curve. This invention eliminates approximation errors at their source, significantly improving adjustment accuracy and linear continuity; it can simultaneously correct long and short wave irregularities in one go, greatly improving operational efficiency; and through iteration, it converges the deviations, meeting millimeter-level high smoothness standards, and has strong versatility.

This invention belongs to the field of building deviation detection technology, and particularly relates to a method for detecting building settlement and displacement deviation based on visual processing. The method first selects natural and artificial landmarks with long-term stability, and after AI visual stability classification, constructs a global benchmark landmark library by combining it with absolute geodetic coordinates. It then acquires multi-view sequential images of the target building and landmarks through a mobile visual acquisition platform, constructs a unified global absolute coordinate system using the benchmark landmarks as global constraints, and generates a high-precision 3D digital model of the target building with absolute 3D coordinates. The model is repeatedly acquired and modeled at a preset cycle, and multi-cycle model registration is completed using stable landmarks as anchor points. Differential calculations are then used to solve the building's full-dimensional deformation deviation data, completing compliance judgment and graded early warning. This invention solves the pain points of traditional detection benchmarks being prone to failure and incomplete detection dimensions, significantly improving the accuracy and timeliness of building deformation detection and early warning.

This invention discloses a point cloud registration device and method based on laser dot matrix features. The device includes a multi-degree-of-freedom motion device, a support platform, a dot matrix laser emitter, a take-up and put-down drive mechanism, and a control and processing module. The core of the method is to control the laser emitter to switch between an avoidance position and a working position through the take-up and put-down drive mechanism, avoiding occlusion of the field of view while projecting a laser dot matrix onto the workpiece surface as a stable artificial feature. A two-level strategy combining deep learning and sub-pixel localization is employed to extract the 3D coordinates of the feature points with high precision. Using these feature points, combined with robot kinematics and pose graph optimization, high-precision and robust point cloud registration is achieved. This invention is particularly suitable for high-precision 3D reconstruction of workpieces with missing surface features, effectively improving the automation and accuracy of registration.

This utility model discloses an auxiliary disk for collecting measurement control points in metal mines, belonging to the field of metal mine measurement and scanning technology. It includes a base with measurement control point holes on the upper part and fixing holes and bolts on the lower part. This utility model aligns the handle of a handheld 3D laser scanner with the auxiliary disk, uses the fixing bolts to secure it to the handheld 3D laser scanner handle through the fixing holes on the lower part of the base, and aligns the measurement control point holes on the upper part of the base with the metal mine measurement control points. This enables the handheld 3D laser scanner to directly, quickly, and accurately collect the 3D coordinate data of the metal mine measurement control points.

This invention discloses a 3D human pose estimation method, device, and storage medium, belonging to the field of computer vision technology. To address the problems of existing technologies ignoring inconsistencies in human body part motion, susceptibility to noise interference from auxiliary views, and disruption of spatiotemporal entanglement structures, this invention proposes a cross-view feature adaptive enhancement and spatiotemporal collaborative modeling network. First, multi-view 2D keypoints are acquired; then, enhanced pose features are generated by jointly modeling the overall human body and part-specific features through a hierarchical knowledge extraction module; auxiliary view features are extracted and filtered using the spatial correlation of the current view as weights, achieving cross-view adaptive enhancement and fusion; subsequently, dual normalization along the channel and time dimensions is performed, and global spatiotemporal correlation is calculated by combining spatiotemporal identifier encoding to extract spatiotemporal collaborative features that retain the original motion entanglement structure; finally, the 3D coordinates are regressed and output. This invention effectively improves the estimation accuracy and robustness of 3D human pose.

An entry-level scanning system includes a first scanner device coupled to a support structure and has a first projector, a first tracking camera and first TOF cameras, while a second scanner device is coupled to the support structure and includes a second projector, a second tracking camera and a second ToF camera, and a processor determines a position and orientation of the system along with 3D coordinates of an object in an environment using image data and 3D coordinate data acquired by either or both of the first scanner device and the second scanner device. The first ToF camera(s) have a first vertical orientation and the second ToF cameras have a second vertical orientation such that the first ToF cameras and the second ToF cameras have a cumulative vertical field of view of about 180 degrees.

This invention discloses a method and apparatus for 3D human pose estimation based on multimodal and occlusion perception, relating to the field of computer vision. The method includes: S1, acquiring a multimodal video sequence including RGB, depth, and infrared sequences; S2, preprocessing to obtain a data-aligned multimodal video sequence; S3, constructing a 3D human pose estimation model including a spatiotemporal feature extraction network, an occlusion detection network, a temporal coordinate classification module, a Z-axis depth prediction module, and a 3D coordinate stitching module; S4, training the 3D human pose estimation model using the data-aligned multimodal video sequence; and S5, using the trained model to perform 3D human pose estimation. This invention effectively handles the 3D human pose estimation problem in complex occlusion scenarios by fusing RGB, depth, and infrared three-modality data, using an occlusion-aware temporal coordinate classification method to predict X and Y axis coordinates, and employing Z-axis depth prediction based on binning and multimodal multi-resolution probabilistic fusion.

This invention discloses a method for planning machining paths of variable ratio gears based on tooth profile point cloud feature extraction, relating to the field of CNC machining technology. This method utilizes an imaging system to acquire imaging images of the variable ratio gear tooth profile surface, and obtains accurate variable ratio gear tooth profile point cloud data through pixel matrix processing and 3D coordinate transformation. After preprocessing, feature information of the variable ratio gear tooth profile is extracted, and finally, a tool machining path is generated based on this feature information. This method, through accurate point cloud acquisition and feature extraction, improves the matching degree between the machining path and the actual tooth profile and machining process, reduces machining deviations, and effectively improves the accuracy and efficiency of variable ratio gear machining, meeting the machining requirements of commercial vehicle steering systems.

This invention discloses a composite tracked triangular wheel omnidirectional chassis and a dynamic weighted trajectory planning method, relating to the field of autonomous navigation technology. A visual perception module processes acquired terrain images and obstacle information and inputs them into a perception network, outputting terrain complexity coefficients and obstacle 3D coordinates. A heterogeneous motion constraint model of the composite tracked triangular wheel chassis is constructed to determine kinematic boundaries. A terrain-chassis dual-drive dynamic weighted evolution model is established, introducing coupling factors to dynamically allocate four-dimensional weights. A dynamic velocity window is generated through a two-layer strategy of illegal velocity pruning and artificial potential field obstacle avoidance. The trajectory is predicted based on an improved DWA algorithm, and the trajectory score is calculated using four-dimensional weights. The optimal trajectory is output, and the motion state is fed back to form a closed loop. This invention uses the aforementioned composite tracked triangular wheel omnidirectional chassis and dynamic weighted trajectory planning method to improve the problems of single chassis performance and poor algorithm-hardware compatibility.

A method generates a three-dimensional (3D) line segment from a two-dimensional (2D) image. A 3D coordinate space along with a first plane in the coordinate space are received. An image with an associated camera pose in the 3D coordinate space is also received. A 2D line segment is detected in the image and then projected into the 3D coordinate space based on the camera pose. The projected 2D line segment is intersected with the first plane, which generates a 3D line segment defined by the intersection. By leveraging the camera pose and first plane, the 3D line segment can be generated from a single 2D image, eliminating the need to triangulate line segments across stereo image pairs.

This invention relates to the field of intelligent monitoring technology, specifically a method for locating personnel trajectories in a bidding evaluation base based on multiple cameras and facial recognition. The method includes: acquiring real-time video streams from various areas of the bidding evaluation base; binding identity tags to areas where facial recognition is successful; converting facial pixel coordinates to 3D world coordinates using camera intrinsic parameters and pose to obtain instantaneous position points; associating consecutive timestamp position points of the same identity to form a single-camera original sequence; based on the spatial layout and overlapping fields of view of multiple cameras, fusing sequences from different cameras of the same identity, and deduplicating and optimizing to generate a global position point sequence; forming a smooth sequence through time-aligned interpolation; calculating displacement vectors and combining them with map corrections to generate personnel movement trajectories. This method achieves high-precision positioning of personnel trajectories in a bidding evaluation base through accurate conversion from 2D pixels to 3D coordinates and optimization of position point fusion from overlapping fields of view of multiple cameras.

This invention relates to a method for reconstructing and predicting the 3D trajectory of a tennis ball based on monocular video. The method includes: matching key points on a 2D court with corresponding points on a 3D standard court to calculate the camera's intrinsic and extrinsic parameters, establishing a 3D world coordinate system based on the 3D standard court; using the known 3D coordinates of racket key points, combined with their 2D coordinates, calculating the racket's 3D depth and pose within the monocular video frame to determine the position of the hitting point in 3D space; and setting hitting point and landing point constraints, adjusting the input initial velocity vector and spin vector, and using physical motion equations to construct a simulated 3D trajectory passing through the hitting and landing points, then projecting the simulated 3D trajectory onto a 2D plane and fitting it with a 2D pixel trajectory to construct the optimal 3D trajectory. This method can reconstruct the 3D trajectory and output the ball's true velocity, net clearance height, and landing depth, meeting the needs of professional quantitative analysis.

This invention relates to the field of construction safety monitoring technology, and discloses a construction site safety early warning system based on UWB and AI cameras. Through multi-source sensor calibration and synchronization, it collects UWB ranging and video frames, and outputs aligned multi-source data. Based on time-frequency domain features and residual networks, it identifies non-line-of-sight locations and compensates for deviations, combining adaptive unscented Kalman filtering to output worker 3D coordinates. It detects targets and behaviors in video frames, achieving cross-modal binding through trajectory constraints and depth metric learning. It analyzes the BIM model to obtain hazardous zone boundaries, using instance segmentation and point cloud reconstruction for correction, and sets dynamic warning zones based on location covariance. It outputs tiered warnings through three-level risk assessment and attention fusion. Edge nodes trigger alarms locally, and edge-cloud collaboratively synchronizes data to the cloud, outputting records and reports. This invention can achieve multi-modal fusion perception and tiered early warning of worker location, behavior, and hazardous zones in complex environments such as tunnel construction.

The invention discloses a 3D machine vision device without an active light source. A system is composed of a left camera, a right camera, an image large model reasoning unit and a 3D machine vision unit. The implementation method of the 3D machine vision device is as follows: 1) the left camera images a measured object, and the object recognition unit recognizes the measured object; (2) imaging the measured object by the right camera, identifying the measured object by an image large model reasoning unit, and determining that the measured object and the measured object of the left camera are the same object; 3) the left camera determines a unique feature pixel in the image of the measured object; (4) the right camera searches for pixels matched with the unique feature pixels of the left camera in the image near the base line; 5) the 3D machine vision unit calculates the parallax formed by the unique feature pixel on the left camera and the right camera, and further calculates the distance value and the 3D point cloud of the measured object; and 6) completing 3D reconstruction of the detected object based on the 3D model of the large model detected object by using the 3D coordinate of the unique feature pixel and the 2D shape of the detected object. The method has the advantages that an active light source is not needed, the detected object in the complex environment is recognized on the left camera and the right camera through image large model reasoning, and the distance of the detected object can be measured in the outdoor environment with strong ambient light to calculate the 3D point cloud.

This invention provides a method, apparatus, and device for removing gross errors in LiDAR-Camera image matching. The method includes: acquiring image data; performing feature extraction and feature matching on the camera images to obtain initial matching point pairs between the camera images and constructing an initial matching point pair set; performing distortion correction processing and ray direction calculation on the image point coordinates of the initial matching point pairs according to the camera imaging model parameters to obtain a normalized ray direction vector; performing 3D coordinate inverse calculation and distance consistency screening on the initial matching point pairs based on the distance map and the normalized ray direction vector to obtain an intermediate matching point pair set; and performing geometric consistency verification on the intermediate matching point pair set based on the prior pose information of the camera images. This invention improves the overall accuracy of LiDAR-Camera data alignment and solves the problems of poor stability and computational efficiency in existing related technologies.