68 results about "Rotation matrix" patented technology

Template points associated with a template are obtained. An image of the template is captured using a camera. The template points of the template are mapped to image points in the image of the template to determine a 2-dimensional transformation matrix (H2D). Pose transform is obtained based on an initial projection matrix resulting from a prior calibration of the camera and the 2-dimensional transformation matrix (H2D), the pose transform having an incomplete rotation matrix. The pose transform is normalized to obtain an updated pose transform having a complete rotation matrix. An affine transform is applied using the complete rotation matrix and a translation vector of the normalized pose transform to the initial projection matrix to obtain an updated projection matrix.

Embodiments of the present application provide a rendering processing method and an electronic device, and relate to the technical field of terminals. The method comprises: determining a first human body model corresponding to each of at least one image frame in a first video. According to inertial measurement unit (IMU) data of a terminal device when the image frame is collected, a rotation matrix corresponding to the image frame is obtained, the rotation matrix being used to indicate rotation of a ground reference coordinate system relative to a device coordinate system of the terminal device. According to the rotation matrix, a first coordinate axis corresponding to a gravity direction in an image coordinate system of the image frame is subjected to rotation processing to obtain a mapped coordinate axis after rotation. Then, the first human body model is subjected to multi-angle rendering processing around the mapped coordinate axis to obtain a rendering result corresponding to the image frame. In this way, the multi-view rendering result of the three-dimensional human body model is rotated around the correct gravity direction, so as to avoid the rendering result of the three-dimensional human body model from being inclined.

This invention discloses a video tracking and recognition method for the spatial trajectory of hoisted objects based on multi-source data fusion, belonging to the field of machine vision monitoring technology. The method includes acquiring pre-loaded physical boundary constants of the tower crane and simultaneously collecting multi-source data during tower crane operation. Based on the physical boundary constants, kinematic calculations are performed on the multi-source data to obtain the relative vertical displacement, the three-dimensional absolute spatial coordinates of the luffing trolley center point, and the absolute coordinates and three-dimensional rotation matrix of the camera's optical center in the global three-dimensional coordinate system. Edge detection is performed on the temporally aligned color video frame matrix to extract candidate wire rope contours and generate a binary edge pixel matrix. Collinear edges are mapped to polar coordinate space using Hough transform. This invention reduces the uncertainty of monocular vision depth scale and improves stability and accuracy under conditions of large-amplitude lifting and complex swinging.

This invention provides a satellite onboard orbit conversion method and system. The method includes: calculating the corrected Julian day at the current time and calculating the Greenwich Mean Sidereal Time Angle based on the corrected Julian day; converting the Greenwich Mean Sidereal Time Angle into radians; calculating the Earth rotation matrix from the J2000 system to the WGS84 system based on the converted radians; pre-calculating the Earth nutation and precession correction matrices as modifiable parameters for satellite commands and updating them periodically; calculating the conversion matrix from the J2000 system to the WGS84 system based on the Earth nutation and precession correction matrices and the Earth rotation matrix; and calculating the satellite position and velocity in the WGS84 system based on the conversion matrix, thereby realizing the satellite onboard orbit conversion. This invention addresses the need for converting EP satellite onboard orbit information from an inertial frame to the WGS84 coordinate system, meets the requirements for short message stand-alone operation, and can provide a reference for on-orbit conversion of other satellites.

The application discloses a robot hand-eye error compensation method, a compensation system and a robot. The method comprises the following steps: performing hand-eye calibration on the robot to obtain a standard hand-eye matrix, which comprises a standard rotation matrix and a standard translation vector; running an end execution module of the robot to a preset introspection observation position, and identifying a reference mark fixedly installed in a field of view to determine a reference center space coordinate and a reference Euler angle corresponding to the reference mark; according to a preset rule, triggering the following introspection compensation operation: running the end execution module to the introspection observation position to determine a current center space coordinate and a current Euler angle; calculating a three-dimensional translation error vector according to the center space coordinate; calculating an Euler angle deviation vector according to the Euler angle; weighting and compensating the standard translation vector by using the translation error vector to obtain a new translation vector, and weighting and compensating the standard rotation matrix by using the Euler angle deviation vector to obtain a new rotation matrix, so as to update the standard hand-eye matrix.

This invention discloses a distributed processing and fusion method and system for electromagnetic signals in multi-source apertures, relating to the field of electromagnetic signal processing, and addressing the problems of pose deflection of detection nodes and low accuracy of multi-frequency signal fusion. First, a moving sensing node is controlled to construct an LC parallel resonant circuit to acquire multi-frequency analog data and simultaneously generate multi-frequency digital voltage characteristic parameters. Then, an inverse projection transformation is performed on the inductive components using a direction cosine space rotation matrix to correct attitude deviations and extract the set of absolute spatial magnetic field vectors in the global spatial coordinate system. Next, the circuit quality factor is extracted as a weighting coefficient to perform superposition fusion, generating a single synthetic target magnetic field vector. Finally, the synthetic vector is correlated with the displacement sequence to perform trajectory association and position reconstruction analysis, outputting a three-dimensional coordinate dataset of the target linear radiation source. This invention achieves high-precision attitude compensation and signal purification, improving the accuracy of target localization in complex and confined spaces.

This invention discloses a stepwise collaborative calibration method for hand-eye extrinsic parameters of a robotic arm in large-scale 3D reconstruction. It acquires 3D point cloud data from multiple observation poses and simultaneously generates initial hand-eye extrinsic parameters. Based on geometric consistency constraints of point-to-surface residuals, the hand-eye extrinsic parameters are iteratively optimized. In each iteration, the point cloud is transformed to the robot's base coordinate system based on the current hand-eye extrinsic parameters and matched, resulting in a set of matched point pairs in the robot's base coordinate system. A point-to-surface residual based on consistency geometric constraints is constructed using this set of matched point pairs as the target. The translation vector and rotation matrix are updated alternately. Finally, the optimal hand-eye extrinsic parameters are selected from those obtained in each optimization round. This invention constructs geometric consistency constraints based on point-to-surface residuals and performs block-based alternating updates of translation and rotation components to achieve posterior compensation for residual errors in the hand-eye extrinsic parameters, thereby improving the consistency and stability of initial registration of multi-view point clouds.

The application discloses a polar coordinate decoupling-based image mirroring and rotating unified transformation method, which comprises the following steps: S10, inputting parameters: an original image, a transformation center point coordinate, a mirroring direction parameter and a rotating angle; S20, traversing each pixel position of an output image; S30, unified mirroring and rotating transformation mapping; S40, interpolation sampling; S50, assigning to the output image; S60, traversing all pixels until completion; and S70, outputting the image. The application abandons the traditional linear algebra idea of firstly constructing a mirroring matrix and then multiplying the mirroring matrix with a rotating matrix, and instead reconsiders the combined operation of mirroring and rotating from the perspective of polar coordinate transformation. In a local polar coordinate system around an arbitrary specified center point, the mirroring operation can be regarded as a sign operation or a shift operation on a polar angle, and the rotating is an addition operation on the polar angle. The two can be combined into a simple expression in the polar angle domain.

The application provides an optimization method and device of a hash encoder, electronic equipment and a storage medium, and belongs to the technical field of vector coding. The method comprises the following steps: determining a dimension importance score according to the information entropy and variance of each vector dimension of training data, selecting N key dimensions with the highest importance after eliminating highly relevant redundant dimensions to construct a rotation matrix, and configuring a plurality of pre-training hash encoders; determining a target hash encoder, encoding input data, and monitoring the coding quality in real time; when quality drift is detected, triggering a parameter optimization mechanism, adjusting the configuration parameters of the target hash encoder in multiple optimization rounds; until the optimization target is reached, applying the optimized target configuration parameters; when the incremental threshold is reached, constructing a rotation matrix increment based on the N key dimensions of the new data, weighting and fusing the current rotation matrix and the rotation matrix increment to obtain an updated rotation matrix, and updating the target hash encoder. By using the application, adaptive parameter optimization can be realized.

This invention relates to the field of infrastructure engineering technology, specifically to a machine vision-based method and system for locating spraying areas. The method includes: using a depth camera to capture the original point cloud of the beam surface, segmenting and selecting target planes on the beam surface using the RANSAC algorithm; converting the image to two dimensions, extracting line features using Hough transform and edge detection, inversely mapping back to three dimensions, and selecting the uppermost point of the line; moving the camera upwards to capture the point cloud of the upper part of the beam, removing interference to extract two target planes; calculating the normal vectors of the two planes, comparing the point cloud to obtain the coordinates of the calibration points, constructing a new three-dimensional coordinate system and generating a rotation matrix; and planning the robot spraying path based on the beam surface design parameters. The system includes: a depth camera, point cloud segmentation, two-dimensional mapping and feature extraction, three-dimensional coordinate recovery and key point selection, spatial calibration, and path planning modules, which work together to implement the above method, eliminate point cloud blind spots, improve positioning accuracy, and meet the requirements for precise spraying of beam surfaces.

This invention discloses a 3D reconstruction method, apparatus, unmanned device, and storage medium. The method includes: acquiring real-time position data of the unmanned device during flight and images captured by a camera; transforming the real-time position data to obtain real-time position data in a first coordinate system; performing 3D reconstruction based on the images and the real-time position data in the first coordinate system to obtain 3D point cloud data in the first coordinate system; acquiring a target vector in the first coordinate system, first pose data of the camera, and first point cloud coordinates of the 3D point cloud data; calculating a rotation matrix based on the first pose data, first point cloud coordinates, and target vector; and transforming the 3D point cloud data based on the rotation matrix to obtain a 3D reconstructed model. This invention improves the accuracy of 3D reconstruction and solves the technical problem of low accuracy in existing 3D reconstruction technologies.

This invention provides a multi-view 3D pose reconstruction method, comprising: constructing a virtual simulation environment; setting the parameters of the virtual simulation environment using a domain randomization strategy; acquiring multi-view images and extracting 2D joint coordinates and confidence scores; normalizing the 2D joint coordinates based on the parameters of each virtual camera to obtain normalized coordinates; concatenating the normalized coordinates with the confidence scores and inputting them into a shared-weight attention encoder; learning the joint topology within a single view through self-attention and outputting view features; concatenating the rotation matrix and translation vector of each view and mapping them to an explicit extrinsic semantic vector via a multilayer perceptron; inputting the view features and extrinsic semantic vector together into a global attention module; performing geometric consistency evaluation through an attention mechanism; and outputting 3D joint coordinates. This method solves the problems of weak cross-view generalization ability, large interference from differences in camera intrinsic parameters, and insufficient robustness against occlusion in existing technologies for multi-view 3D pose estimation.

The application provides a method for converting robot coordinate systems to a three-dimensional coordinate system and a virtual platform. The method comprises: obtaining more than one robot uploaded reference point coordinates and to-be-converted point coordinates in corresponding robot coordinate systems; creating an auxiliary coordinate system in a three-dimensional coordinate system of the virtual platform; converting the to-be-converted point coordinates of each robot into coordinates of the to-be-converted point in the three-dimensional coordinate system, which comprises: corresponding the reference point coordinates in the robot coordinate system to the auxiliary coordinate system; obtaining a comprehensive rotation matrix and a position offset by using the reference point coordinates in the auxiliary coordinate system and the reference point coordinates in the three-dimensional coordinate system; multiplying the to-be-converted point coordinates by the comprehensive rotation matrix, and adding the position offset to the multiplied coordinates to obtain the coordinates of the to-be-converted point in the three-dimensional coordinate system. The application also provides a virtual platform, which comprises: a receiving module, a display module and a robot three-dimensional coordinate acquisition module. With the aid of the auxiliary coordinate system and the reference point, the conversion process is simplified, and the calculation amount is reduced.

The invention provides an attitude determination method and device, which can be applied to the technical field of attitude measurement. The attitude determination method comprises the following steps: acquiring initial attitude data of a cooperative target under an inertial coordinate system by using an inertial measurement unit; processing the initial attitude data based on a rotation matrix to obtain first rotation attitude data under a target coordinate system, the rotation matrix being used for representing a relationship between the attitude data under the inertial coordinate system and the attitude data under the target coordinate system, and the target coordinate system being a coordinate system constructed based on a cooperative target; under the condition of receiving visual attitude data output by the visual measurement unit, fusing the first rotating attitude data with the visual attitude data to obtain first fused attitude data; the visual attitude data is attitude data obtained by performing data acquisition on the cooperative target while the inertial measurement unit acquires the initial attitude data; and resolving the first fusion attitude data, and determining first target attitude data of the cooperative target.

The application discloses a kind of double plane observation constraint baseline direction's active vision calibration method, it is related to optical active vision field.The application includes the following steps: constructing optical active vision system, including emitting device and imaging device, the external parameter between emitting device and imaging device is described by rotation matrix and baseline vector;The emitting device and imaging device are placed in reference plane of different depth, the plane homography of image derivation is carried out to different depth reference plane, and plane constraint is constructed, and the unknown baseline direction is connected with observation homography and plane parameter;Plane constraint relation couples baseline vector b with the plane parameters of two different depth reference planes, and the rank-one decomposition of scaled homography difference is obtained, to obtain secondary constraint;Physical feasible root is selected, and baseline direction can be directly obtained.The robustness and flexibility of the application make it suitable for real-world active vision system.

The application provides a kind of relative attitude autonomous measurement method between multiple space vehicles based on space-based radar precision measurement.Space formation space vehicles are all equipped with space-based radar, and can measure relative distance and angle, which is converted into LOS observation vector in each space vehicle body coordinate system;Since the same physical line-of-sight vector in different space vehicle body coordinate systems satisfies the rotation matrix transformation relationship, multiple non-collinear LOS vectors are obtained through mutual observation between multiple space vehicles, and a least squares optimization problem is constructed;The optimal rotation matrix is solved by singular value decomposition, and the result is constrained to SO (3) group to ensure the legality of the rotation matrix, and the complete three-axis relative attitude between space vehicles is finally solved.The application does not need to know the relative attitude and position of the space vehicle as priori, and can realize the relative attitude measurement between any two space vehicles through distance and angle measurement values.

The application relates to an intelligent ship pilotage method and system based on RTK and inertial navigation fusion, which comprises the following steps: deploying and calibrating an RTK base station at a high point of a wharf, generating a differential correction quantity in real time and broadcasting the differential correction quantity to a shipborne terminal; collecting shore feature points by using an RTK surveying and mapping device, constructing a continuous and smooth three-dimensional shore curve model by using a cubic spline interpolation algorithm, and generating a shore visual map with geographic coordinates by fusing wharf geographic information; establishing a ship coordinate system, converting inertial navigation attitude data to the coordinate system by using a rotation matrix and a translation vector, fusing the RTK position data, and outputting high-precision ship position and posture; transmitting the fused data through a 5G network, converting the data to a local plane coordinate system at a shore terminal, generating a real-time dynamic visual interface, pushing the interface to a tugboat terminal through an RTSP protocol, and assisting in pilotage operation, so that high-precision perception, visual guidance and shore-ship cooperation in a ship berthing process are realized, and the safety and intelligent level of pilotage are improved.

The application discloses a kind of based on motion compensation time domain filtering method for image, comprising: using autoregressive moving average model to the obtained gyroscope error is modeled, the change rule of data is used to estimate model parameter, and gyroscope error model is established;Using the gyroscope data of removing error obtained by gyroscope error model, the rotation matrix of camera is calculated, and the rotation image is compensated in reverse using rotation matrix, and the de-rotation image is obtained.Through kalman filtering to image motion vector sequence, image random jitter is obtained, and the de-rotation image is compensated according to random jitter, and the stable image and the subjective motion vector of image are obtained;Using the motion target detection algorithm in difference map based on large map coordinate system to detect target in stable image.The application has high detection sensitivity, and can detect target and its trajectory quickly through a small amount of frame track correlation.

A method, apparatus, device, and readable storage medium for extracting interface center coordinates are disclosed, relating to the field of industrial inspection technology. The method includes filtering initial 3D point cloud data corresponding to a target interface to obtain target point cloud data; performing principal component analysis on the target point cloud data to determine the target tilt direction; calculating the target tilt angle based on the angle between the target tilt direction and a horizontal reference plane, and constructing a rotation matrix based on the target tilt angle; rotating the target point cloud data using the rotation matrix to generate a horizontal attitude point cloud; and performing geometric fitting on the horizontal attitude point cloud to output the target interface center coordinates. This application enables precise correction of the interface attitude, improving the accuracy of interface center coordinate extraction.

This invention discloses a method and system for smoothing quantization of large models based on learnable rotation matrix Kronecker integral solutions, relating to the field of artificial intelligence model compression technology. The method includes: constructing a search space containing Kronecker integral solutions of different numbers of matrices, with the tensor-matrix i-th modal product as the computational implementation method for each solution; adaptively selecting the optimal rotation matrix hybrid decomposition scheme for the attention modules and feedforward networks of each decoder layer of the large model through a preset scoring function and layer-by-layer greedy heuristic search; training the complete parameters of the decomposition matrix, learnable weights, activation value truncation threshold, and diagonal scaling matrix parameters of the scheme; fusing the extra parameter matrix on the weight side of the linear layer, retaining the extra parameter matrix on the activation value side, and quantizing the model based on the trained quantization parameters to generate a deployable model. This effectively solves the problems of accuracy loss and computational overhead under low-bit quantization of large models, significantly improving computational and storage efficiency while maintaining high accuracy.

The present application relates to the field of photogrammetry, and discloses a land space image analysis method based on photogrammetry, comprising: acquiring aerial images and attitude data of a survey area to generate sparse point clouds; distinguishing rigid and non-rigid feature points based on semantic segmentation results; establishing a local vertical reference system for non-rigid points and configuring anisotropic parameters with transverse variance greater than longitudinal variance; mapping the parameters to a global weight matrix using a coordinate rotation matrix, and substituting them into a bundle adjustment model to complete iterative solution, the present application effectively realizes the automatic suppression of non-rigid disturbance of vegetation and water area by constructing an anisotropic weight determination mechanism coupled with physical properties and geometric constraints, and improves the geometric accuracy and stability of three-dimensional reconstruction in complex land space scenes.

The application discloses a double-layer nonlinear optimization monocular vision hand-eye calibration method based on outlying sample screening. h2e An Euler angle transformation is carried out to ensure the orthogonality of the rotation matrix R; an optimization model H1 of the hand-eye matrix is constructed, and a traditional LM nonlinear optimization method is used to perform first optimization on the initial value of the hand-eye matrix to obtain a hand-eye matrix after the first optimization; the optimization model H1 of the hand-eye matrix is modified to obtain a second optimization model H2, and then an LM nonlinear optimization method with the outlying sample screening method is used to perform second optimization on the hand-eye matrix after the first optimization to obtain a hand-eye matrix after the second optimization; and the method can reduce the influence of random errors existing in the monocular vision system of the robot and improve the calibration precision of the hand-eye matrix of the robot.

This invention discloses a ship attitude measurement method based on a quadrangular frustum infrared sensor array. The method includes: installing a quadrangular frustum base on a top platform directly above the ship's center of gravity, ensuring its bottom surface is parallel to the XOY plane of the ship's coordinate system; arranging a set of mid- and long-wave infrared sensors on each of the four sides, with the planes containing the bow-stern and port-starboard sides being orthogonal to each other; during measurement, obtaining the corresponding normal vectors from the four sides and constructing two sets of orthogonal sensor coordinate systems using the right-hand rule; based on the mapping relationship between the sensor coordinate system and the ship's coordinate system, deriving the mathematical expression of the sensor-observed target in the ship's coordinate system through rotation matrix transformation, thus realizing the ship's attitude angle calculation. This invention proposes a multi-directional infrared-assisted ship attitude measurement method with a standardized structure, clear coordinate relationships, and easy calibration. It is highly scalable and can provide high-precision, robust attitude measurement for ships in complex sea conditions.

This invention discloses a method for local feature transfer and fusion of a 3D model, belonging to the field of 3D reconstruction technology. The method includes: loading and parsing the original model and the target model; constructing and caching a mesh topology adjacency list to achieve O(1) complexity neighbor lookup; selecting the corresponding boundary control circle and performing centroid normalization; calculating the optimal rotation matrix through singular value decomposition of the covariance matrix, and constructing a rigid transformation matrix by combining an adaptive scaling factor; replacing the vertices of the local region of the transformed target model with those of the original model to generate a fused model; performing multi-level expansion of the splicing region based on the adjacency list, and applying Taubin forward and reverse smoothing iterations to eliminate seams. This invention ensures registration accuracy and global optimality through SVD decomposition, effectively prevents model volume shrinkage through Taubin bidirectional smoothing, and significantly improves computational efficiency through the adjacency list caching mechanism, achieving high-quality, high-efficiency, and automated local feature transfer and fusion of a 3D model.

The application discloses a binocular camera secondary calibration method and device, the method comprises the following steps: taking two target targets by a camera to be calibrated; extracting left target corner points of the two target targets in a left image respectively, and extracting right target corner points of the two target targets in a right image respectively; extracting a left vanishing point in the left image according to the left target corner points of the two target targets, and extracting a right vanishing point in the right image according to the right target corner points of the two target targets; calculating a pitch angle and a yaw angle of a rotation matrix between the left eye camera and the right eye camera according to the left vanishing point and the right vanishing point; calculating left three-dimensional coordinates and right three-dimensional coordinates of the target corner points based on a monocular principle; and calculating a translation matrix between the left eye camera and the right eye camera according to the left three-dimensional coordinates and the right three-dimensional coordinates. In the automatic driving or assisted driving scene, the calculation accuracy of the rotation matrix and the translation matrix is improved, so that the accuracy of the binocular camera secondary calibration is improved.