53 results about "Stereo reconstruction" patented technology

A system and method for extracting structure from stereo that represents the scene as a collection of planar layers. Each layer optimally has an explicit 3D plane equation, a colored image with per-pixel opacity, and a per-pixel depth value relative to the plane. Initial estimates of the layers are recovered using techniques from parametric motion estimation. The combination of a global model (the plane) with a local correction to it (the per-pixel relative depth value) imposes enough local consistency to allow the recovery of shape in both textured and untextured regions.

The present invention belongs to the field of computer-aided human face image re-construction technology. The present invention features that human face as model establishing object is first photographed with two, one upper and one lower, stere ovideo cameras to obtain video sequence from the front to the side of the face with unchanged expression; the precise posture parameters in different times are then obtained via autoamtic posture estimating algorithm; after the initiation of model and posture, 3D information are extracted separately from two types of 2D threads including one stereo reconstruction type of the image and the other matching and correcting type of model projection contour and image contour; and the 3D face model is finally established.

The invention discloses a three-dimensional reconstruction method for a concrete CT (computed tomography) image, comprising the following steps of: firstly, performing tomography on a concrete test sample by a CT machine; secondly, performing three-dimensional reconstruction on the concrete CT image by using a ray-casting algorithm, so as to obtain real images of aggregates, mortar and holes; and finally, performing three-dimensional stereo reconstruction on the concrete CT image by using a visualization tool VTK (visualization tool kit). The three-dimensional reconstruction method for the concrete CT (computed tomography) image, disclosed by the invention, has the advantages of being strong in real-time performance, high in operation efficiency, and capable of clearly and reliably distinguishing the aggregates, the mortar and the holes. A real meso-structure model of concrete can be obtained through a three-dimensional reconstruction result, the three-dimensional reconstruction result can be subjected to a test comparison with a concrete mesoscopic numerical simulation result, and a numerical analysis model is corrected.

The invention provides a stereo-hyperspectral human face recognition method based on auroral imaging, for solving the technical problems of poor recognition result stability and low recognition efficiency of conventional multi-spectral human face recognition methods. The stereo-hyperspectral human face recognition method comprises the following steps of: 1, using an auroral spectrum system to collect human face Information; 2, carrying out stereo reconstruction of a human face according to a two-dimensional auroral image in the human face information; 3, diving the two-dimensional human face into regional blocks, and dividing the three-dimensional human face and human face spectrum information into regional blocks according to the division result; 4, extracting comprehensive feature vectors of the regional blocks; 5, for many human faces and the to-be-recognized human face, adopting the step 1 to the step 4 to obtain sample library data and to-be-recognized human face data; 6, training the sample library data to get a classifier; 7, inputting the to-be-recognized human face data to the classifier, and identifying the human face. The stereo-hyperspectral human face recognition method of the invention has the characteristics of good stability and high identification efficiency, and can be used to authenticate an identity for financial transaction, security defense, customs clearance and the like.

The embodiment of the invention provides a method for reconstructing a three-dimensional model of high dynamic range texture, comprising the following steps: adopting shooting devices arranged in a grouped encircling array mode to shoot multi-viewpoint and multiple exposed images and / or videos; restoring response curves of the shooting devices; performing exposure alignment on the images shot at different exposure time so as to restore the three-dimensional model based on the response curves of the devices; and restoring high dynamic texture information by calculating color information of each vertex of a grid. The embodiment of the invention also discloses a device for reconstructing the three-dimensional model of high dynamic range texture, which is composed of an acquisition module, a response curve restoring module, an exposure alignment module and a reconstructing module. The method and the device for reconstructing the three-dimensional model of high dynamic range texture can realize multi-viewpoint stereo reconstruction for the three-dimensional model of high dynamic range under the premise of not lowering geometric reconstruction precision of the model without additionally updating the hardware equipment.

A novel stereo reconstruction pipeline that features depth map alignment and outlier identification is provided. One example method includes obtaining a plurality of images depicting a scene. The method includes determining a pose for each of the plurality of images. The method includes determining a depth map for each of the plurality of images such that a plurality of depth maps are determined. Each of the plurality of depth maps describes a plurality of points in three-dimensional space that correspond to objects in the scene. The method includes aligning the plurality of depth maps by transforming one or more of the plurality of depth maps so as to improve an alignment between the plurality of depth maps. The method includes identifying one or more outlying points. The method includes generating a three-dimensional model of the scene based at least in part on the plurality of depth maps.

The invention relates to a multi-view stereo reconstruction method based on an Internet photo gallery which utilizes a plurality of photos of the same scene at different viewing angles obtained from the Internet so as to restore the sparse three-dimensional point cloud of the scene for obtaining camera parameter, and then densify the sparse three-dimensional point cloud for restoring the dense three-dimensional point cloud of the scene. The method provided by the invention comprises the following specific steps that the geometry and movement at the scene are stored at the same time, that is, the camera is demarcated and the sparse three-dimensional point cloud of the scene is restored; image screening is performed, a nearest image collection is selected for each photo, and an optimal image sub collection is selected for each three-dimensional point; the three-dimensional points are densified so as to count out new three-dimensional points, and an area generation method of taking the conventional three-dimensional points as a center and generating towards the near area is adopted; and the larger-error three-dimensional points are filtered and got rid of. The invention provides a photo screening method, avoids redundant calculation, upgrades the speed of model generation, filters and gets rid of inaccurate three-dimensional points, and improves the accuracy of the model generation.

The invention discloses a quantization coding method and a quantization coding device of a prediction residual signal. The method comprises the following steps of: acquiring two paths of signals according to left and right sound channel signals, respectively converting the two paths of signals to a frequency domain, and carrying out prediction residual coding by using the two paths of signals of the frequency domain to obtain the prediction residual signal; calculating characteristic parameters of the left and right sound channel signals in the frequency domain according to the two paths of signals; and comparing the characteristic parameters with set threshold values, selecting different quantization algorithms according to a comparison result, and carrying out quantization coding on theprediction residual signal. By the embodiment of the invention, a reconstructed stereo signal can provide better balance between the stability of a sound field and an expanded spectral range, and better stereo reconstruction quality is obtained.

A computer-implemented method of performing a combined surface reconstruction and registration of stereo laparoscopic images includes a computer system generating an intraoperative three-dimensional model of an anatomical area of interest and receiving a plurality of stereo endoscopic images of the anatomical area of interest from a laparoscope. The computer system performs a first stereo reconstruction process using the stereo endoscopic images to yield a first surface image corresponding to the anatomical area of interest and aligns the first surface image with a surface of the intraoperative three-dimensional model using a rigid registration process. Then, the computer system performs a second stereo reconstruction process using the first surface image and prior knowledge from a pre-operative three dimensional model to yield a second surface image corresponding to the anatomical area of interest and aligns the second surface image with the surface of the intraoperative three-dimensional model using a non-rigid registration process.

In an embodiment, a method determines a three-dimensional model from a plurality of images taken of a geographic region by one or more cameras from different perspectives. The method includes determining, using a first stereo reconstruction technique: (i) a plurality of three-dimensional candidate surface points from the plurality of images, and (ii) which of the images in the plurality of images view each of the plurality of candidate surface points. The method also includes identifying an empty space between each of the plurality of candidate surface points and each camera model for the respective images determined to view the candidate surface point. The method further includes for each of a plurality of pairs of images from the plurality of images, determining, using a second stereo reconstruction technique, a surface estimate for the pair of images. The method also includes merging the surface estimates to identify a final surface.

The invention discloses a semantic stereo reconstruction method for a remote sensing image, and mainly solves the problem of low semantic three-dimensional reconstruction precision caused by ignoringrelated information of semantic segmentation and parallax estimation in the prior art. According to the implementation scheme, firstly, experimental data are preprocessed; a semantic segmentation network and a parallax estimation network are trained by using the training data; the trained network is tested on the test image, and test results of different frequency band information are fused to obtain a fused semantic segmentation result and a fused parallax result; the error correction module is used for assisting each other to correct the error part of the opposite side; and the disparity information is calculated to obtain height information, and the semantic segmentation result is combined with the height information to obtain a semantic three-dimensional reconstruction result of the image. According to the method, the proportion of small samples is improved, the influence of data on the network is balanced, the semantic information and the parallax result are fused with each other,the accuracy of semantic three-dimensional reconstruction of the remote sensing image is improved, and the method can be used for urban scene three-dimensional reconstruction.

The invention discloses a stereo matching method of three-dimensional reconstruction. The method comprises the steps that first, the accumulation phenomenon of Harris image feature points is improved, the defect of low pixel accuracy is overcome, inhibiting factors are introduced to enable the image feature points to be evenly distributed, and then a Gauss curved surface fitting and interpolating method is used for enabling the acquired image feature points to achieve a sub-pixel level; then SIFT is used for describing the image feature points to generate a feature vector for matching, the problem that the Harris image feature points are sensitive to image gradation nonlinear transformation is solved, then a random sampling algorithm RANSAC with good robustness is used for estimating a fundamental matrix, at the same time, mismatching point pairs are further removed, and at last under guide of the fundamental matrix, convection matching is conducted to search for unmatched image feature point pairs which possibly reflect object structures. The stereo matching method of three-dimensional reconstruction enables the surface structure of objects to be better restored, and is effectively used for non-calibration image information three-dimensional reconstruction.

Stereo image reconstruction techniques are described. An image from a root viewpoint is translated to an image from another viewpoint. Homography fitting is used to translate the image between viewpoints. Inverse compositional image alignment is used to determine a homography matrix and determine a pixel in the translated image.

According to some embodiments of the present invention, a system for stereo reconstruction from a monoscopic endoscope is provided. The monoscopic endoscope comprising an image pick-up element at a distal end thereof and a working channel defined by a body of the monoscopic endoscope. The working channel provides a port at the distal end of the monoscopic endoscope. The system for stereo reconstruction comprises a light patterning component configured to be disposed within the working channel of the monoscopic endoscope such that a light emitting end of the light patterning component will be fixed with a defined relative distance from the distal end of the image pick-up element. The system for stereo reconstruction also includes a data processor adapted to be in communication with the image pick-up element. The light patterning component forms a pattern of light that is projected onto a region of interest. The data processor is configured to receive an image signal of the region of interest that includes the pattern, and determine a distance from the endoscope to the region of interest based on the image signal and based on the defined relative distance between the light emitting end of the light patterning component and the distal end of the image pick-up element.

The invention provides an underwater binocular stereo vision measuring system and method based on a dynamic reference coordinate system. A dynamic reference coordinate system is built through using four active cursor targets to eliminate a measurement error caused by vibration. An underwater stereo vision reconstruction method based on a light field model is used to calculate a three-dimensional coordinate of a target angular point under a scanning system. A coded point coordinate system is established by carrying out singular value decomposition on a coordinate matrix of the target angular point. A spatial three-dimensional point of a laser stripe in a scanning system coordinate system is calculated by using an underwater laser stripe matching and stereo reconstruction method based on the light field model. The measurement error caused by the vibration is eliminated by calculating the three-dimensional coordinate of a laser stripe center point of each group of pictures under the scanning system and converting to the scanning system coordinate system of the first group of picture target points. An underwater target point is packaged in a high pressure resistant state and can withstand high pressure of deep sea 4000 meters. Therefore, the system can be used for a dynamic high precision measurement scene of the deep sea 4000 meters.

In this invention, systems and methods are described that allow performing stereo reconstruction with line-scan cameras in general configurations. Consequently, the cumbersome exact alignment of the sensor lines becomes superfluous. The proposed method requires that telecentric lenses instead of perspective lenses are mounted on the line-scan cameras. In this case, the images can be accurately rectified with stereo rectification. The rectified images can be used to perform an efficient stereo matching. The method comprises a camera model and a calibration procedure that allow to precisely model the imaging process, also including the modelling of lens distortions even if the sensor lines are not exactly mounted behind the principal points. This ensures a high accuracy of the resulting 3D reconstruction.

The invention provides a DSM generation method and device based on unmanned aerial vehicle images, and the method comprises the steps of obtaining an unmanned aerial vehicle image set, and determining the corresponding camera initial internal parameters according to the label data of the unmanned aerial vehicle images in the unmanned aerial vehicle image set; determining an affine transformation relationship between a reference image and a search image according to the feature points corresponding to homonymous image points in an overlapping region of the reference image and the search image in the unmanned aerial vehicle image set; determining the three-dimensional coordinates of the feature points corresponding to the homonymy image points in the overlapping region of the image of which the overlapping region is greater than a preset threshold value, and generating a DSM sparse point cloud according to the three-dimensional coordinates and the initial internal parameters; and obtaining a DSM dense point cloud according to the sparse point cloud by using a multi-view stereo reconstruction method, and generating a DSM corresponding to the unmanned aerial vehicle image. Therefore, the disordered data set can be effectively utilized to carry out rapid and accurate adjustment calculation to realize the three-dimensional reconstruction of a large scene, and the corresponding DSM is generated.

The invention discloses a remote sensing image binocular stereo matching method based on a deep network and semantic information, and mainly solves the problem of low stereo matching precision of an existing remote sensing image. The method comprises the following steps: 1) collecting US3D data set remote sensing images; 2) using the training semantic segmentation network model to obtain semanticprior of the remote sensing image; 3) carrying out stereo matching on the remote sensing image by using a deep network model; 4) carrying out stereo matching on the remote sensing image by using a traditional SGBM algorithm; 5) fusing the matching results of the step 3) and the step 4) by using semantic prior to obtain a fused parallax image; 6) denoising the fused parallax graph, and 7) correcting the denoised graph in the step 6) by using semantic priori to obtain a binocular stereo matching result. Compared with the prior art, the binocular stereo matching precision of the remote sensing image is improved, and the binocular stereo matching method can be used for three-dimensional space stereo reconstruction of the remote sensing image.

The invention discloses an auxiliary robot for total knee joint replacement surgery. Two groups of stereo vision and light source mechanical arms with seven joints are designed, and the function of rapid and accurate stereo reconstruction of knee joint surgery is achieved; two mechanical arms with supporting and clamping positioning components are designed, the fixation of thighs and cruses in thetotal knee joint replacement surgery is achieved, and thus knee joints are fixed, and guarantee is provided for the accuracy of the surgery; two mechanical arms for turnover clamping of skin and muscle tissues are designed, and four novel turnover clamping components work together to realize minimally invasive and accurate turnover of skin and muscle tissues of a knee joint incision and reduce secondary trauma of the surgery incision turnover; an surgery mechanical arm with four executive components is designed, the knee joint skin and muscle tissue cutting, incision hemostasis, leg bone cutting, leg bone drilling and other functions for the knee joint replacement surgery are achieved, real-time function switching is achieved, and the accuracy, stability and efficiency of the surgery canbe greatly improved.

An improved vehicle surface scanning system for assessing the damaged surfaces of a vehicle with resulting estimates of repair costs. A mobile scanning booth is assembled in an open-ended tunnel-like rig having a plurality of reflective panels positioned along opposite sides and across the roof of the booth to serve as deflection screens. A plurality of scanner modules are mounted in fixed positions about opposite ends of the booth and positioned to face the interior of the booth. Wheels provide controlled locomotion / movement of the scanning booth over the vehicle. The scanner modules use a combined hybrid methodology of active stereo 3D reconstruction and deflectometry to acquire data measurements along the surfaces of the vehicle incrementally as the booth is moved. The incremental measurement data acquired during the mobile scanning is processed and furthermore combined to produce accurate reports of the damage surfaces and estimates of associated repair costs.

The invention relates to an audio generation method and device, computer equipment and a storage medium. The method can be applied to an application scene for generating an audio in a map navigation process, and comprises the following steps: based on a distance value between a sound source and a target object, performing distance sensing processing on an audio signal of the sound source to obtain a distance sensing audio signal; performing sound mixing processing on the distance sensing audio signals of the sound sources in the same grid area to obtain sound mixing audio signals corresponding to each grid area; determining a sound image center based on the position information and the audio weight value of each sound source in the grid area; based on the sound image center, carrying out stereo reconstruction on the sound mixing audio signal of the grid area, and acquiring a stereo audio signal; and respectively carrying out sound mixing processing on the left channel audio signals and the right channel audio signals of all the stereo audio signals to obtain stereo sound mixing audio signals. By adopting the method, the calculation overhead for generating the stereo audio signal can be effectively reduced.

A hybrid 3D optical scanning system combining methodologies of active stereo 3D reconstruction and deflectometry to provide accurate 3D surface measurements of an object under inspection. In a preferred embodiment, the present system comprises a calibrated digital camera stereo pair and first digital projector for active stereo 3D reconstruction metrology positioned at selected distance apart on the same horizontal plane facing the inspection object at an optimal incidence angle relative to the object surface. A secondary digital projector for deflectometry is directed to project a pattern sequence across the optical projection path of the first digital projector and onto a nonplanar deflection screen extended between the digital camera and the inspection object at a selected angle relative thereto, the selected angle being chosen so that that the pattern projected onto the deflection screen by the secondary digital projector reflects from the surface of the inspection object towards the digital camera. Calibrated active stereo 3D reconstruction is used to derive global shape and the uncalibrated nonplanar deflected light acquires curvature of local fine deformations and surface imperfections along the object surface. The calibrated camera allows per pixel correlations to be defined between the separate acquisition methods and produces accurate absolute measurements of both shape and local features of specular, glossy, metallic curved surfaces.

The invention discloses a multi-eye stereo imaging microscope and an imaging method thereof. The multi-eye stereo imaging microscope includes a lens barrel structure, a microscope main body supportingstructure, and an electric objective table. The imaging method of the multi-eye stereo imaging microscope comprises the following steps of: (1) shooting a checkerboard calibration board, and using atwo-step calibration method to perform a camera calibration on multi-eye CCD electronic eyepieces to determine the relative position of each eyepiece; (2), performing image registration on a capturedsample according to the camera parameters obtained in step (1), so as to prepare for stereo reconstruction; (3) performing feature point stereo matching on the sample by using a improved microscope-specific matching algorithm SLM-SGBM; and (4) performing parallax optimization on a disparity map obtained in step (3) by using an optimization algorithm to obtain a final stereo matching result. The invention can realize real-time three-dimensional reconstruction of the surface of a minute sample when only using the multi-eye stereo imaging microscope.

This disclosure is related to a non-contact tool center point calibration method for a robot arm, and the method comprises: obtaining a coordinate transformation relationship between a flange surfaceof the robot arm and cameras by a hand-eye calibration algorithm; constructing a space coordinate system by a stereoscopic reconstruction method; actuating a replaceable member fixed with the flange surface to present postures in a union field of view of the cameras sequentially, recording feature coordinates of the replaceable member in the space coordinate system, and recording flange surface coordinates which is under the postures in the space coordinate system; obtaining a transformation relationship between a tool center point and the flange surface; and updating the transformation relationship into a control program of the robot arm. Moreover, the disclosure further discloses a calibration device performing the calibration method and a robot arm system having the calibration function.

The invention discloses a face recognition method based on a pole imaging stereo camera and face recognition equipment. The method comprises the steps that the stereo camera based on a pole imaging system is constructed; the stereo camera is utilized to collect face data to obtain a two-dimensional pole image of a face; affine transformation is performed on the two-dimensional pole image to obtain images of the face at three different viewpoints; a multi-viewpoint stereo matching method is utilized to perform stereo reconstruction on the images of the face at the three different viewpoints to obtain a three-dimensional structure of the face; the face in the two-dimensional pole image is detected and partitioned to obtain a complete two-dimensional face image; key points and local features of the face are extracted according to the complete two-dimensional face image and the three-dimensional structure of the face, and feature vectors are calculated to obtain depth information and feature vector information of the face; multiple pieces of face information is collected to establish a face database; and to-be-recognized face information is compared with the information in the face information base to judge whether the to-be-recognized face information is consistent with the information in the face information base. Through the method, the face information can be recognized easily, quickly and accurately.

The invention provides an image three-dimensional reconstruction method for enhancing real vision. The method comprises the following steps: acquiring an original image; performing image segmentation on the original image according to a preset edge detection technology, and combining a preset image edge splicing technology to obtain a wheel gallery graph; adopting a preset affine transformation matrix to calculate the vestibule map to obtain a three-dimensional imaging target image, and utilizing the obtained target image to perform three-dimensional modeling to obtain a three-dimensional model; and fusing the original image and the obtained three-dimensional model, and combining the target image to obtain a three-dimensional visual image. By implementing the method, interactive image stereoscopic vision virtual simulation generation can be realized on the basis of a space model generation technology of image three-dimensional affine transformation, so that the problems of high reconstruction cost, low robustness, low efficiency and the like of an existing stereoscopic vision image are solved.