37 results about "Camera matrix" patented technology

The invention belongs to the technical field of augmented reality (XR) and computer vision crossing, relates to fusion application in immersive digital performance, and provides a virtual actor reconstruction and immersive presentation system based on 4D Gaussian splash modeling and XR space positioning. The system comprises a set of spherical multi-camera-position high-synchronization camera shooting matrix used for capturing dynamic images of actors; performing dynamic modeling on the multi-angle image through a 4D Gaussian splashing technology, and outputting a virtual actor point cloud model which can be used by XR equipment; vPS visual positioning and an SLAM tracking module are combined, and precise mapping positioning of a performance space is achieved in AR / MR equipment. The method supports the immersive watching of the actor image at the audience end at a 360-degree free visual angle, and realizes the natural presentation of the virtual actor without dead angles and wearing in cooperation with shielding judgment and a real-time rendering engine. The system is widely applicable to on-site entertainment scenes such as immersive theaters, text travel performances, brand activities, concerts and television programs.

The invention discloses a multi-view mark-point-free facial expression capturing method and system, which are applied to the technical field of facial expression capturing, and the method comprises the steps: collecting a human face multi-view image based on a multi-dimensional camera matrix; performing two-dimensional key point detection and neutral expression face three-dimensional reconstruction on the face multi-view image to obtain a standard face topology model; acquiring a non-neutral expression depth map from the face multi-view image; optimizing a model coefficient based on the standard face topology model, the non-neutral expression depth map and the face multi-view image; and inputting the optimized model coefficient into a standard face topology model to obtain a real face facial expression model. According to the invention, the accuracy and integrity of the acquired image can be improved through the acquired multi-view-angle image of the human face; the expression base coefficient, the translation coefficient and the Euler angle rotation coefficient are optimized, so that a more accurate and more real facial expression model is obtained based on a standard facial topology model, and the accuracy of expressions and poses is ensured.

The invention discloses a single-image three-dimensional head reconstruction method and system based on a guide diffusion model. The method comprises the following steps: obtaining a head portrait from a single image to be reconstructed; reconstructing a three-dimensional head model according to the head portrait; performing weak projection mapping on the three-dimensional head model under the predicted camera matrix to the head portrait, and obtaining a three-dimensional head model with front textures by utilizing pixel information of the head portrait; performing cylindrical surface UV expansion on the three-dimensional head model with the front texture to obtain a two-dimensional UV texture map and recording a UV mapping relation; inputting the two-dimensional UV texture map into the guided diffusion network model to predict the texture of the invisible area, and repairing to obtain a complete two-dimensional UV texture map; and then the complete two-dimensional UV texture image is reversely pasted to the three-dimensional head model. According to the method, the three-dimensional head model which is high in fidelity, free of artifacts and rich and vivid in texture details can be reconstructed from individual single portrait images of different ages, genders, races and facial expressions.

The invention discloses a monocular vision 3D target labeling method based on multi-modal data, and relates to the technical field of image processing, and the method comprises the steps: obtaining a 2D data set generated for a training region; determining camera external parameters corresponding to the camera generating the 2D data set, and determining a camera matrix corresponding to the camera in combination with the camera external parameters; determining a projection matrix of the training model in combination with the auxiliary information file corresponding to each RGB image, and training a 3D target detection model based on the projection matrix in combination with all the RGB images, the coordinate position of the 2D target detection frame corresponding to each RGB image and the camera matrix; and obtaining 2D image data collected for the to-be-detected area, and outputting a 3D detection frame corresponding to each piece of 2D image data in the trained 3D target detection model. In combination with the projection matrix, the network model can learn the perspective relation and angular point information of the target in the image, parameter convergence of the neural network is helped, and an ideal 3D detection frame prediction result is achieved.

The present application discloses a method, device, and medium for collecting information on overlapping screens. The method includes: collecting information displayed on a screen through a camera matrix to obtain several partial images corresponding to the screen; in the case of the information displayed on the screen corresponding to a railway dispatching centralized control system, obtaining a standard image corresponding to the screen, and obtaining the coordinate information of the indicator lights on the screen through the standard image; determining the relative position relationship of each camera in the camera matrix, and selecting a designated indicator light as a reference point within a preset edge area of ​​several partial images; splicing the several partial images according to the relative position relationship and the coordinate information of the reference point to obtain a spliced ​​image, and monitoring the operating status of the railway dispatching centralized control system through the spliced ​​image. The method achieves the acquisition of the display content of the screen to be collected without accessing the system.

The utility model discloses a multi-camera image acquisition device for building a three-dimensional model of a bridge structure prefabricated part, which comprises a portal steel frame and a track assembly, and the portal steel frame comprises truss columns respectively positioned on two sides of the bridge structure prefabricated part and a truss beam positioned above the bridge structure prefabricated part. The truss girder is mounted at the tops of the two truss columns; cameras are arranged on the truss columns and the truss girders at intervals to form a camera matrix which is fixed in relative position and is used for acquiring image data of the bridge structure prefabricated parts; the rail assembly comprises sliding rails arranged corresponding to the two truss columns respectively, the sliding rails are parallel to the length direction of the bridge structure prefabricated part, and the truss columns are installed on the corresponding sliding rails in a sliding fit mode. According to the multi-camera image acquisition device for building the three-dimensional model of the prefabricated part of the bridge structure, image data acquired by the cameras can be fused by fixing the position relation between the cameras, so that the three-dimensional model of the prefabricated part of the bridge structure is built.

The present invention discloses a monocular vision 3D target labeling method based on multimodal data, which relates to the field of image processing technology. The method comprises: obtaining a 2D data set generated for a training area; determining the camera extrinsic parameters corresponding to the camera generating the 2D data set, and determining the camera matrix corresponding to the camera in combination with the camera extrinsic parameters; determining the projection matrix of the training model in combination with the auxiliary information file corresponding to each RGB image, and training a 3D target detection model based on the projection matrix, in combination with all RGB images, the coordinate position of the 2D target detection frame corresponding to each RGB image, and the camera matrix; obtaining 2D image data collected for the area to be detected, and outputting the 3D detection frame corresponding to each 2D image data in combination with the trained 3D target detection model. The present invention, combined with the projection matrix, can enable the network model to learn the perspective relationship and corner point information of the target in the image, help the parameters of the neural network converge, and achieve ideal 3D detection frame prediction results.

The invention provides a texture mapping method and system for a three-dimensional model and terminal equipment, and the method comprises the steps: obtaining an original three-dimensional model, and processing the original three-dimensional model to obtain a simplified three-dimensional model; obtaining a camera internal parameter matrix and a camera external parameter matrix, wherein the camera external parameter matrix is pose information of different shooting angles corresponding to the original three-dimensional model; determining a first corresponding relation matrix between a pixel coordinate system of a physical shooting image corresponding to the original three-dimensional model and a coordinate system of the original three-dimensional model based on the camera external reference matrix and the camera internal reference matrix; and determining a two-dimensional texture image of each shooting angle based on the first corresponding relation matrix and the original three-dimensional model, and determining a second corresponding relation between the two-dimensional texture image and the grid vertex of the simplified three-dimensional model, so as to carry out texture mapping on the simplified three-dimensional model based on the second corresponding relation. The problem that in the prior art, the image shadow is mapped to the model, so that the color of the model is different and unreal is solved.

This application provides a method, electronic device, and storage medium for object implantation in multi-degree-of-freedom (DOF) video. The method includes: obtaining the coordinates of a real camera matrix in a real-world coordinate system, wherein the real camera matrix is ​​used to capture multi-DOF video; constructing a virtual coordinate system based on the coordinates of the real camera matrix in the real-world coordinate system, wherein the virtual coordinate system is the coordinate system corresponding to the virtual camera matrix in virtual space; mapping the trajectory information of the real camera matrix in the real-world coordinate system to the trajectory information of the virtual camera matrix in the virtual coordinate system; and associating the trajectory information of the virtual camera matrix in the virtual coordinate system with a virtual object to be implanted in the multi-DOF video to obtain implantation information for implanting the multi-DOF video. This application provides a foundation for accurate matching between virtual objects and multi-DOF video.

The invention provides an unmanned aerial vehicle automatic inspection route generation method and system based on visual identification, and relates to the technical field of unmanned aerial vehicle inspection, and the method comprises the steps: collecting an inspection video stream of an inspection region through a camera device carried by an inspection unmanned aerial vehicle; extracting an inspection image frame and carrying out target detection on the inspection image frame based on a deep learning model to generate a multi-scene inspection target point; based on a camera matrix and a digital elevation model, pixel coordinates of the multi-scene inspection target point are converted into actual geographic coordinates; automatically planning a structured inspection route of the inspection unmanned aerial vehicle according to the actual geographic coordinates in combination with a path curvature constraint condition; a structured inspection route is converted into a recognizable route instruction file and is imported into a flight control system of the inspection unmanned aerial vehicle, so that high efficiency, precision, automation and intelligence of the whole inspection process of the unmanned aerial vehicle are realized by fusing visual recognition and intelligent route planning technologies, and the system adaptability, the inspection quality and the operation safety are greatly improved.

An image processing device 10 comprises: a camera matrix acquisition unit 11 that acquires a camera matrix for each image when point group data of an object is generated from a plurality of images; a projection vector calculation unit 12 that selects a pair of images from the plurality of images and calculates a projection vector from one of the images in the pair to the other image by using the camera matrix of each of the images in the pair; and a feature-point pair setting unit 13 that selects a feature point in one of the images in the pair, sets a projection area, in which the selected feature point is projected in the other image, by using the selected feature point and the projection vector, specifies a feature point corresponding to the selected feature point from among feature points existing in the projection area, and sets the two feature points as a feature point pair.

The application discloses a kind of intelligent detection building health and risk prediction teaching system, belong to intelligent construction teaching field, including: through intelligent car and the high-definition camera matrix of unmanned aerial vehicle terminal carrying is collected building data and is transmitted to operation super brain;Through the AI recognition big model and structure safety, natural disaster, personnel safety three risk assessment sub-models built-in in operation super brain, the multi-level comprehensive risk assessment of building state is carried out;Through the hazard identified in building digital twin 3D model corresponding position visualization presentation and generating graded warning signal by wisdom management platform, platform supports student login operation to carry out picture digitization, autonomous operation practice and complete experiment report;After completing on-campus laboratory simulation and course practice, form course outline and experiment report and popularize to other colleges and universities.The application enables students to visually observe building hazards, complete risk prediction operations, and promote cross-campus sharing of teaching resources.

A method for creating an image of a base of a container uses a matrix camera. The matrix camera captures a series of individual captures of regions of the base of the container which overlap in sections, while the base of the container is transilluminated by a light source arranged on the side opposite the matrix camera. A digital image of the base of the container is compiled from the individual captures. The invention is characterized in that the region of the base of the container captured in an individual capture is defined by an illumination structure, that the illumination structure is shifted relative to the base of the container between two individual captures, and that the container and the matrix camera remain rotationally invariant to one another during the capturing of the series of individual captures. The invention also relates to a device for carrying out the method.

The invention discloses an intelligent cockpit gesture recognition method fusing attention and map updating, and the method comprises the steps: building a space model of a cockpit, presetting a local attention interval for each seat of the cockpit, associating the local attention interval with a multi-camera matrix, and constructing a seat-interval-camera mapping relation; the method comprises the following steps: synchronously acquiring image data in a cabin through a multi-camera matrix, operating a hand detection model to detect a hand image, and preliminarily classifying the hand image into a local attention interval; performing gesture recognition on the hand image, and outputting a gesture recognition result; building a cockpit knowledge graph, and dynamically updating a node and edge relationship of the cockpit knowledge graph according to the passenger state change and / or the gesture recognition result; and performing authority verification and conflict arbitration based on a gesture recognition result and the cabin knowledge graph, sending a control instruction passing arbitration to a corresponding cabin actuator, and performing closed-loop optimization based on user feedback. According to the invention, the accuracy, real-time performance and system robustness of gesture recognition can be improved.

The application discloses a kind of methods for detecting and repairing non-ferrous metal strip surface crack synchronously online, comprising the following steps: system initialization: after system power on, a series of self-checking and calibration operations are performed to ensure that all components are in standby state and initial position; online detection: drive equipment drives the rotation of the reel, and the industrial camera matrix arranged on the lower side of the industrial camera mounting bracket detects the surface crack of the strip passing below; when the online detection result is "crack found", the synchronous mechanism is connected to the main shaft of the drive equipment through the reversing device. Through the multi-level automatic process of detection-repair-reinspection, the traditional mode of relying on manual slow inspection and manual operation during shutdown is replaced. The strip can complete defect identification and processing during high-speed operation without the need to stop or slow down, greatly improving production efficiency while reducing labor intensity, eye strength and skill requirements of workers.

The application discloses a single-image three-dimensional head reconstruction method and system based on a guided diffusion model, and the method comprises the following steps: obtaining a head portrait from a single image to be reconstructed; reconstructing a three-dimensional head model according to the head portrait; weakly projecting and mapping the three-dimensional head model to the head portrait under a predicted camera matrix, and obtaining a three-dimensional head model with front texture by using pixel information of the head portrait; performing cylindrical UV unfolding on the three-dimensional head model with the front texture to obtain a two-dimensional UV texture map and record a UV mapping relationship; inputting the two-dimensional UV texture map into a guided diffusion network model to predict the texture of an invisible area and obtain a complete two-dimensional UV texture map; and pasting the complete two-dimensional UV texture map back to the three-dimensional head model. The application can reconstruct a high-fidelity, artifact-free and realistic three-dimensional head model with rich texture details from a single portrait of an individual of different ages, genders, races and facial expressions.

The invention discloses wide-area intelligent monitoring equipment adopting a medium-long-focus camera matrix, and relates to the technical field of intelligent monitoring equipment.The wide-area intelligent monitoring equipment comprises a monitoring machine head, a bearing mechanism is arranged at the top of the monitoring machine head, and the left side and the right side of the top of the bearing mechanism are connected with a walking mechanism by arranging hinge plates; according to the wide-area intelligent monitoring equipment adopting the medium and long-focus camera matrix, a control module B drives a pushing frame to move through an electric telescopic rod B, the pushing frame drives a pushing rod to move through a connecting frame, the pushing rod drives a contact wheel to rotate, the contact wheel drives a hinged plate to rotate oppositely, and the hinged plate drives a bearing mechanism to move; a control module A drives a power wheel to rotate through an adjusting motor, the power wheel drives a monitoring machine head to rotate through an adjusting rod, meanwhile, an electric hydraulic rod A drives a pull rod to move, the pull rod drives the monitoring machine head to rotate to a proper monitoring angle through a sliding rail, and the monitoring machine head transmits monitored abnormal conditions to a monitoring data platform. The method has the advantage of adaptively adjusting the monitoring angle.

A computer-implemented method for processing anatomical imaging data, the method comprising, for a plurality of anatomical specimens: acquiring 3D imaging data of a particular body part of an anatomical specimen of the plurality of anatomical specimens; acquiring 2D images of the specific body part of the anatomical specimen from a plurality of different perspectives relative to the specific body part; determining a plurality of camera matrixes P corresponding to the obtained 2D image; projecting the acquired 3D imaging data according to the plurality of camera matrixes P to generate a synthesized 2D image; and associating the acquired 2D image with one of the synthesized 2D images as a training data pair according to the plurality of camera matrices P and a plurality of viewing angles. The method further includes generating a training data set for training an artificial intelligence algorithm using the training data pair.

The application discloses a kind of methods for solving camera matrix of monitoring camera in three-dimensional model, comprising: obtaining data and preprocessing;Measure the same point: find at least 5 same points in monitoring camera image and three-dimensional real scene model, and record the image coordinate and object coordinate of same point;Object coordinate is transformed into camera coordinate system, based on the image coordinate of same point and object coordinate under camera coordinate system, and focus three-point collinear, construct equation group;After solving equation group, it is transformed into world coordinate system by coordinate, and output camera matrix.The application solves the problem that existing method is difficult to obtain camera internal parameter and external parameter, and the cost is higher.

The invention discloses a pig measurement method, system and device based on a camera matrix and a medium, and relates to the technical field of animal husbandry informatization. The method comprises the following steps: acquiring a target pig image; processing according to a hair segmentation model and the target pig image to obtain a segmentation image for distinguishing hair and skin contours; the hair noise suppression model is constructed based on a deep learning network; according to the segmented image, a density correction model is used for measurement and evaluation, and a data estimation result of the target pig is obtained; the density correction model is constructed based on a three-dimensional modeling technology and is used for determining the gender corresponding to the pig breed and the body size information and the volume information of the growth stage. The accuracy of pig measurement estimation can be improved.

The invention discloses a method for visually inspecting a camera matrix of a monitoring camera based on a live-action three-dimensional model, and the method comprises the steps: decomposing a camera matrix M to obtain the coordinate of a camera focus F, calculating the photographing light LP of a homonymy point through an image space point P and the camera matrix, obtaining a calculated object space point W'through calculating the intersection point of a straight line LP and a three-dimensional live-action model S, and obtaining the camera matrix of the monitoring camera through the calculated object space point W '. Obtaining a difference vector D according to the real object space point W and the calculated object space point W ', namely the error of the camera matrix; and finally, drawing a vector D in the three-dimensional real scene model S by taking the real object space point W as a starting point, and marking a specific numerical value of the D. The method solves the problems that an existing precision test method of a camera matrix is not fused with a real scene, is not visual enough, and is difficult to obtain an object variance value of errors and distribution characteristics of a three-dimensional space.

The invention relates to the field of image super-resolution, in particular to an image reconstruction method, and particularly relates to a Fourier laminated image reconstruction method based on a deep learning network. A reflective far-field Fourier lamination imaging system is utilized, a series of low-resolution images are shot through a camera array, deep learning is combined with traditional Fourier lamination, the low-resolution images are reconstructed by utilizing a neural network, the acquisition number of the low-resolution images is reduced, the requirement of the traditional Fourier lamination for redundancy is met, and the method has the advantages of high resolution and high resolution. And the sampling speed and the reconstruction speed of Fourier lamination imaging are improved.

Example implementations include a method, apparatus and computer-readable medium for estimating a real world distance in an image, comprising obtaining an image from a camera installed in an environment and identifying, in the image, a first set of parallel lines on a real world ground plane and a second set of parallel lines on the real world ground plane. The implementations include determining a first intersection point of the first set of parallel lines and a second intersection point of the second set of parallel lines, identifying a ground line connecting the first intersection point and the second intersection point, and estimating an internal camera matrix based on a focal length and an image size associated with the camera. The implementations include calibrating the camera for real world distance measurement by determining a perpendicular vector to the ground line based on the internal camera matrix.

The present invention discloses an imaging identification method and system for determining the cone growth region of a KDP crystal, relating to the technical field of KDP crystal quality inspection. The method comprises the steps of obtaining a KDP crystal to be tested and inspecting the crystal; establishing an imaging identification system, and using the imaging identification system to image the KDP crystal to obtain a scattered light image of the KDP crystal. The KDP crystal to be tested is placed in the imaging identification system, and scattered laser light is incident obliquely on the surface of the KDP crystal to be tested, causing the crystal to move horizontally. The scattered light image is captured in real time by a CCD camera matrix; and the crystal growth region at all measurement areas of the KDP crystal to be tested is determined using the scattered light image. The present invention can quickly and effectively use the scattered light image to directly determine the cone growth region of the KDP crystal, resulting in a simpler identification method and more accurate identification results.

1. A computer-implemented method for processing anatomical imaging data, the method comprising: acquiring, for a plurality of anatomical specimens, 3D imaging data of a particular body part of the anatomical specimens among the plurality of anatomical specimens; acquiring 2D images of the particular body part of the anatomical specimens from a plurality of different viewpoints relative to the particular body part; determining a plurality of camera matrices P corresponding to the acquired 2D images; generating a composite 2D image by projecting the acquired 3D imaging data according to the plurality of camera matrices P; and associating the acquired 2D image with one of the composite 2D images as a training data pair according to the plurality of camera matrices P and the plurality of viewpoints. The method further comprises generating a training data set for use in training an artificial intelligence algorithm using the training data pairs.

The invention discloses a method for resolving a camera matrix of a monitoring camera in a three-dimensional model. The method comprises the steps that data are acquired and preprocessed; measuring homonymy points: finding out at least five homonymy points in the monitoring camera image and the three-dimensional real scene model, and recording image space coordinates and object space coordinates of the homonymy points; converting the object space coordinates into image space coordinates based on homonymy points and object space coordinates in a camera coordinate system under the camera coordinate system, and constructing an equation set according to three-point collinearity of a focus; and after solving the equation set, converting coordinates into a world coordinate system, and outputting a camera matrix. The method solves the problems that an existing method is difficult to obtain internal parameters and external parameters of the camera, and the obtaining cost is high.

The invention provides an anti-upheaval method and device for a combined visual system or an enhanced flight visual system, and the method comprises the following steps: determining the position of a real horizon line in a current EVS image, and obtaining the position of the real horizon line on the EVS image through a camera matrix of the EVS, the latitude and longitude of an airplane, the height of the airplane and an airport plane equation. And determining the maximum elevation line position in the current EVS image. Finding an image point with the highest position in the EVS image for elevation points in a given range, and obtaining a straight line which passes through the point and is parallel to the horizon, namely the maximum elevation line. And starting or inhibiting the anti-upheaval function according to the input of the pilot.

A computer-implemented method of estimating a 6D pose and shape of one or more objects from a 2D image, comprises the steps of: detecting, within the 2D image, one or more 2D regions of interest, each 2D region of interest containing a corresponding object among the one of more objects; cropping out a corresponding pixel value array, coordinate tensor, and feature map for each 2D region of interest; concatenating the corresponding pixel value array, coordinate tensor, and feature map for each 2D region of interest; and inferring, for each 2D region of interest, a 4D quaternion describing a rotation of the corresponding object in the 3D rotation group, a 2D centroid, which is a projection of a 3D translation of the corresponding object onto a plane of the 2D image given a camera matrix associated to the 2D image, a distance from a viewpoint of the 2D image to the corresponding object, a size, and a class-specific latent shape vector of the corresponding object.

The present application relates to the field of image super-resolution, more particularly, to an image reconstruction method, in particular to a Fourier stack image reconstruction method based on a deep learning network. By using a reflective far-field Fourier stack imaging system, a series of low-resolution images are captured by a camera array, deep learning is combined with traditional Fourier stack, and a neural network is used to reconstruct low-resolution images, reduce the number of low-resolution image acquisition, reduce the redundancy requirement of traditional Fourier stack, and improve the sampling speed and reconstruction speed of Fourier stack imaging.