Showing papers on "3D reconstruction published in 1996"

Physics-Based Deformable Models: Applications to Computer Vision, Graphics, and Medical Imaging

Abstract: From the Publisher: Physics-Based Deformable Models presents a systematic physics-based framework for modeling rigid, articulated, and deformable objects, their interactions with the physical world, and the estimation of their shape and motion from visual data. This book presents a large variety of methods and associated experiments in computer vision, graphics and medical imaging that help the reader better understand the presented material. In addition, special emphasis has been given to the development of techniques with interactive or close to real-time performance. Physics-Based Deformable Models is suitable as a secondary text for a graduate level course on Computer Graphics, Computational Physics, Computer Vision, Medical Imaging, or Biomedical Engineering. In addition, this book is appropriate as a reference for researchers and practitioners in the above mentioned fields.

Computer-vision-based registration techniques for augmented reality

Abstract: Augmented reality is a term used to describe systems in which computer-generated information is superimposed on top of the real world; for example, through the use of a see-through head-mounted display. A human user of such a system could still see and interact with the real world, but have valuable additional information, such as descriptions of important features or instructions for performing physical tasks, superimposed on the world. For example, the computer could identify objects and overlay them with graphic outlines, labels, and schematics. The graphics are registered to the real-world objects and appear to be “painted” onto those objects. Augmented reality systems can be used to make productivity aids for tasks such as inspection, manufacturing, and navigation. One of the most critical requirements for augmented reality is to recognize and locate real-world objects with respect to the person’s head. Accurate registration is necessary in order to overlay graphics accurately on top of the real-world objects. At the Colorado School of Mines, we have developed a prototype augmented reality system that uses head-mounted cameras and computer vision techniques to accurately register the head to the scene. The current system locates and tracks a set of preplaced passive fiducial targets placed on the real-world objects. The system computes the pose of the objects and displays graphics overlays using a see-through head-mounted display. This paper describes the architecture of the system and outlines the computer vision techniques used.

Euclidean 3D Reconstruction from Image Sequences with Variable Focal Lenghts

Abstract: One of the main problems to obtain a Euclidean 3D reconstruction from multiple views is the calibration of the camera. Explicit calibration is not always practical and has to be repeated regularly. Sometimes it is even impossible (i.e. for pictures taken by an unknown camera of an unknown scene). The second possibility is to do auto-calibration. Here the rigidity of the scene is used to obtain constraints on the camera parameters. Existing approaches of this second strand impose that the camera parameters stay exactly the same between different views. This can be very limiting since it excludes changing the focal length to zoom or focus. The paper describes a reconstruction method that allows to vary the focal length. Instead of using one camera one can also use a stereo rig following similar principles, and in which case also reconstruction from a moving rig becomes possible even for pure translation. Synthetic data were used to see how resistant the algorithm is to noise. The results are satisfactory. Also results for a real scene were convincing.

Accurate Three-Dimensional Reconstruction of Intravascular Ultrasound Data Spatially Correct Three-Dimensional Reconstructions

Abstract: Background The geometrical accuracy of conventional three-dimensional (3D) reconstruction methods for intravascular ultrasound (IVUS) data (coronary and peripheral) is hampered by the inability to register spatial image orientation and by respiratory and cardiac motion. The objective of this work was the development of improved IVUS reconstruction techniques. Methods and Results We developed a 3D position registration method that identifies the spatial coordinates of an in situ IVUS catheter by use of simultaneous ECG-gated biplane digital cinefluoroscopy. To minimize distortion, coordinates underwent pincushion correction and were referenced to a standardized calibration cube. Gated IVUS data were acquired digitally, and the spatial locations of the imaging planes were matched to their corresponding coordinates. Image points were then transformed relative to their respective 3D coordinates, rendered in binary voxel format, resliced, and displayed on an image-processing workstation for off-line analysis. The method was tested by use of phantoms (straight tube, 360° circle, 240° spiral) and an in vitro coronary artery model. In vivo feasibility was assessed in patients who underwent routine interventional coronary procedures accompanied by IVUS evaluation. Actual versus calculated point locations were within 1.0±0.3 mm of each other (n=39). Calculated phantom volumes were within 4% of actual volumes. Phantom 3D reconstruction appropriately demonstrated complex morphology. Initial patient evaluation demonstrated method feasibility as well as errors if respiratory and ECG gating were not used. Conclusions These preliminary data support the use of this new method of 3D reconstruction of vascular structures with use of combined vascular ultrasound data and simultaneous ECG-gated biplane cinefluoroscopy.

Knowledge-directed vision: control, learning, and integration

Abstract: The knowledge-directed approach to image interpretation popular in the 1980's, sought to identify objects in unconstrained two-dimensional (2-D) images and to determine the three-dimensional (3-D) relationships between these objects and the camera by applying large amounts of object- and domain-specific knowledge to the interpretation problem. Among the primary issues faced by these systems were variations among instances of an object class and differences in how object classes were defined in terms of shape, color, function, texture, size, and/or substructures. This paper argues that knowledge-directed vision systems typically failed for two reasons. The first is that the low- and mid-level vision procedures that were relied upon to perform the basic tasks a vision were too immature at the time to support the ambitious interpretation goals of these systems. This problem, we conjecture, has been largely solved by recent advances in the field of 3-D computer vision particularly in stereo and shape reconstruction from multiple views. The other impediment was that the control problem for vision procedures was never properly addressed as an independent problem. This paper reviews the issues confronted by knowledge-directed vision systems, and concludes that inadequate vision procedures and the lack of a control formalism blocked their further development. We then briefly introduce several new projects which, although still in the early stage of development, are addressing the complex control issues that continue to obstruct the development of robust knowledge-directed vision systems.

Affine reconstruction from perspective image pairs with a relative object-camera translation in between

Abstract: A method is described to recover the three-dimensional affine structure of a scene consisting of at least five points identified in two perspective views with a relative object-camera translation in between. When compared to the results for arbitrary stereo views, a more detailed reconstruction is possible using less information. The method presented only assumes that the two images are obtained by identical cameras, but no knowledge about the intrinsic parameters of the camera(s) or about the performed translation is assumed. By the same method, affine 3D reconstruction from a single view can be achieved for parallel structures. In that case, four points suffice for affine reconstruction.

Evaluation of OS-EM vs. ML-EM for 1D, 2D and fully 3D SPECT reconstruction

Abstract: For SPECT reconstruction, iterative Maximum Likelihood Expectation Maximization (ML-EM) estimation has a huge computational burden. The objective of this paper is to compare images obtained by ML-EM and an EM algorithm acting on Ordered Subsets of projections (OS-EM). Two digital phantoms, a cylinder with two cold spots and an ellipsoid with several hot spots and one cold spot were reconstructed from 120 simulated noisy projections. 1D (/spl delta/-like point source response), 2D (single slice response) and fully 3D reconstruction were investigated. Three quantities were calculated for the evaluation, viz. contrast, normalized standard deviation and mean squared error. In the case of fully 3D reconstruction, OS-EM 60 reconstructions (i.e., using 60 ordered subsets) were very close to ML-EM reconstructions. This shows that the OS-EM algorithm is an extremely fast and efficient method to accelerate iterative SPECT reconstruction with speed-up factors of close to half the number of projections.

3D reconstruction of indoor environments

Abstract: This paper presents a new 3D scene analysis system that automatically reconstructs the 3D model of real-world scenes from multiple range images acquired by a laser range finder on board of a mobile robot. The reconstruction is achieved through an integrated procedure including range data acquisition, geometrical feature extraction, planning the next view, registration and integration of multiple views. The system relies only on the acquired data. Experimental results on real range images are presented in the paper. Direct applications of this technique include 3D reconstruction and/or update of architectural or industrial plans into a CAD model, design verification of buildings, navigation of autonomous robots, and input to virtual reality systems.

Fast iterative image reconstruction of 3D PET data

Abstract: For count-limited PET imaging protocols, two different approaches to reducing statistical noise are volume, or 3D, imaging to increase sensitivity, and statistical reconstruction methods to reduce noise propagation. These two approaches have largely been developed independently, likely due to the perception of the large computational demands of iterative 3D reconstruction methods. The authors present results of combining the sensitivity of 3D PET imaging with the noise reduction and reconstruction speed of 2D iterative image reconstruction methods. This combination is made possible by using the recently-developed Fourier rebinning technique (FORE), which accurately and noiselessly rebins 3D PET data into a 2D data set. The resulting 2D sinograms are then reconstructed independently by the ordered-subset EM (OSEM) iterative reconstruction method, although any other 2D reconstruction algorithm could be used. The authors demonstrate significant improvements in image quality for whole-body 3D PET scans by using the FORE+OSEM approach compared with the standard 3D Reprojection (3DRP) algorithm. In addition, the FORE+OSEM approach involves only 2D reconstruction and it therefore requires considerably less reconstruction time than the 3DRP algorithm, or any fully 3D statistical reconstruction algorithm.

From 2D images to 3D face geometry

Abstract: This paper presents a global scheme for 3D face reconstruction and face segmentation into a limited number of analytical patches from stereo images. From a depth map, we generate a 3D model of the face which is iteratively deformed under stereo and shape-from-shading constraints as well as differential features. This model enables us to improve the quality of the depth map, from which we perform the segmentation and the approximation of the surface.

What accuracy for 3D measurements with cameras

Abstract: We estimate the internal and external parameters of the camera, and simultaneously the distortion's parameters. Our aim is the selection of the best distortion model, using statistical test for the importance of distortion parameters. Also we examine the accuracy in camera parameter estimation and 3D reconstruction, in relation with the noise in the image. We answer the question "until which level of noise, is it possible to obtain a good camera parameter estimation, and from there a good reconstruction?". Experiments are evaluated on simulated data and a rest is performed with real data.

Optical 3D motion measurement

Abstract: This paper presents a CCD-camera based system for high-speed and accurate measurement of the three-dimensional movement of reflective targets. These targets are attached to the moving object under study. The system has been developed at TU Delft and consists of specialized hardware for real-time multi-camera image processing at 100 frames/s and software for data acquisition, 3D reconstruction and target tracking. An easy-to-use and flexible but accurate calibration method is employed to calibrate the camera setup. Applications of the system are found in a wide variety of areas; biomechanical motion analysis for research, rehabilitation, sports and ergonomics, motion capture for computer animation and virtual reality, and motion analysis of technical constructions like wind turbines. After a more detailed discussion of the image processing aspects of the system the paper will focus on image modeling and parameter estimation for the purpose of camera calibration and 3D reconstruction. A new calibration method that has recently been developed specifically for the measurement of wind turbine blade movements is. Test measurements show that with a proper calibration of the system a precision of 1:3000 relative to the field of view dimensions can be achieved. Future developments will further improve this result.

REALISE: reconstruction of REALity from Image SEquences

Abstract: REALISE was designed to extract from sequences of images, acquired with a moving camera, the information necessary for determining the 3D (CAD-like) structure of a real-life scene together with information about the radiometric signatures of surfaces bounding the extracted 3D objects (e.g. reflectance behaviour). The retrieved information is then integrated in a virtual reality (VR) software environment. The R&D work is been performed principally in the following areas of computer vision and computer graphics: structure from motion, recovery of geometries, recovery of photometric and texture information, highly realistic rendering on the basis of empirically-based reflectance models, and the design and development of improved rendering processes together with a new VR system. Beside this innovative R&D work another key aspect of REALISE is to have computer vision & computer graphics cooperate to produce realistic 3D data efficiently.

Using 2D active contour models for 3D reconstruction from serial sections

Abstract: Reconstructing the 3D boundary of a structure from tomographic data is an important problem in medical imaging. Various forms of the active surface model have been proposed to tackle it, but they are either computationally expensive or requiring extensive guidance from the user. We propose an active model that is conformable to any 3D shape, easy to use, efficient, and is highly parallelizable. It requires only an initial contour on one of the tomographs to start with. The essential idea is to break down the highly interconnected 3D deformation system into 2D pieces that work independently to save computations, and that deform in all directions to wrap around noise spots while conforming to the structure's shape.

Euclidean 3D reconstruction from stereo sequences with variable focal lengths

Abstract: A stereo rig can be calibrated using a calibration grid, but recent work demonstrated the possibility of auto-calibration. There remain two important limitations, however. First, the focal lengths of the cameras should remain fixed, thereby excluding zooming or focusing. Second, the stereo rig must not purely translate, which however is the most natural type of motion. This also implies that these methods collapse when the motion comes close to being a translation.

Camera viewpoint control for the automatic reconstruction of 3D objects

Abstract: An algorithm for the image dependent control of the camera viewpoint is presented, which is applied to the automatic reconstruction of 3D objects. For computer animation applications, "shape from silhouettes" using equally distributed viewpoints is an often used reconstruction technique. With respect to the local reconstruction errors, the use of equally distributed camera views is unfavourable for arbitrary shaped objects. For that reason, a camera viewpoint control is introduced, which purposefully rotates a turntable with the 3D object depending on the trace of the silhouette contour points over the rotation angle. This trace provides information about the location of object planes and gives a measure for the expected local 3D reconstruction errors. It turns out, that the new algorithm reduces the remaining 3D reconstruction errors up to 70% compared to algorithms without viewpoint control using the same number of viewpoints.

Data Fusion in the Field of Non Destructive Testing

Abstract: This paper deals with the reconstruction of metallic blocks from gam- magraphic and ultrasonic data. The inverse problems that have to be faced are solved within a Bayesian framework. Firstly, the problem of 3D reconstruction from gammagraphies is discussed. Then, we propose a new deconvolution method for ultrasonic traces. Finally, we present a reconstruction method that accounts for both sets of data: the ultrasonic data is used to detect breaks in the object and those bounds are then incorporated in the prior model for the reconstruction from gammagraphies.

Automated 360° profilometry of a three-dimensional diffuse object and its reconstruction by use of the shading model.

Abstract: The 360° profilometry of a three-dimensional (3-D) diffuse object by use of the light intersection and its image reconstruction by surface shading are presented. The lack of data in one direction, which was due to occlusion, was compensated by the projection of two lines of light from different directions. Some experiments to profile objects and their reconstruction by computer are shown. The entire surface model was constructed, and a real shading image was obtained by means of computer graphics.

Computer assisted coronary intervention: 3D reconstruction and determination of optimal views

Abstract: The authors propose a novel method for computer assisted coronary intervention. The 3D coronary arterial tree is first reconstructed based on a pair of angiograms acquired from single-plane or bi-plane systems at arbitrary orientations. With this 3D structure, the views of selected vessel segments with minimal vessel foreshortening and overlapping are determined in the form of a sequence of gantry angulations. A computer simulation confirmed the accuracy of 3D centerline reconstruction to within 2.5% error. To date 3D reconstruction have been completed in 20 RCA, 16 LCA, and 3 bypass grafts. With the proposed technique, coronary sizing can be optimized for lesion length. Practice interventions can be completed in this computer assisted process that should lead to more accurate, efficient, and safe coronary procedure.

Active shape inferring based on the symmetry in stable poses-shape from function approach

Abstract: Automatic acquisition of a three-dimensional object model is one of major goals of computer vision research. We develop a new algorithm for reconstructing the 3D shapes of objects from multiview images which are automatically acquired based on the symmetry in stable poses. Our approach is based on a "shape from function" strategy. That is, shape is inferred by observing and estimating "the stability in stable poses", which is the most fundamental condition for objects to be functioning in the real world. Based on that assumption, we first observe the object from the vertical top and estimate a plane of symmetry using the center of gravity and the symmetrical axis of the silhouette in the top view image. We then estimate the next viewpoints, from which the estimated symmetry are examined with silhouettes obtained as projections on the estimated plane of symmetry. By executing this symmetry estimation-examination process recursively, we reconstruct the 3D shape of the object as a product of straight homogeneous generalized cylinders (SHGCs) whose cross section and the axis direction are constrained from the silhouette at each view-point and the viewing direction respectively by using octree representation or a CAD system. Experimental results show that the effectiveness of the approach in reconstructing the 3D shape of object.

Accurate 3-D reconstruction from trinocular views through integration of improved edge-matching and area-matching techniques

Abstract: This paper describes a method for obtaining a reliable 3D reconstruction of close-range objects by properly combining edge- and area-based matching techniques. The adopted acquisition system is a set of three calibrated low-cost CCD cameras. By using an accurate camera model and camera calibration, the method is capable of working with any camera setup. The proposed technique has been tested on some real scenes with encouraging results. Some of these experimental results are presented here.

Three-dimensional reconstruction under varying constraints on camera geometry for robotic navigation scenarios

Abstract: 3D reconstruction is an important research area in computer vision. With the wide spectrum of camera geometry constraints, a general solution is still open. In this dissertation, the topic of 3D reconstruction is addressed under several special constraints on camera geometry, and the 3D reconstruction techniques developed under these constraints have been applied to a robotic navigation scenario. The robotic navigation problems addressed include automatic camera calibration, visual servoing for navigation control, obstacle detection, and 3D model acquisition and extension. The problem of visual servoing control is investigated under the assumption of a structured environment where parallel path boundaries exist. A visual servoing control algorithm has been developed based on geometric variables extracted from this structured environment. This algorithm has been used for both automatic camera calibration and navigation servoing control. Close to real time performance is achieved. The problem of qualitative and quantitative obstacle detection is addressed with a proposal of three algorithms. The first two are purely qualitative in the sense that they only return yes/no answers. The third is quantitative in that it recovers height information for all the points in the scene. Three different constraints on camera geometry are employed. The first algorithm assumes known relative pose between cameras; the second algorithm is based on completely unknown camera relative pose; the third algorithm assumes partial calibration. Experimental results are presented for real and simulated data, and the performance of the three algorithms under different noise levels are compared in simulation. Finally the problem of model acquisition and extension is studied by proposing a 3D reconstruction algorithm using homography mapping. It is shown that given four coplanar correspondences, 3D structures can be recovered up to two solutions and with only one uniform scale factor, which is the distance from the camera center to the 3D plane formed by the four 3D points corresponding to the given four correspondences in the two camera planes. It is also shown that this algorithm is optimal in terms of the number of minimum required correspondences and in terms of the assumption of internal calibration.

Integrated real-time motion segmentation and 3D interpretation

Abstract: This paper describes a real-time integrated motion segmentation and 3D reconstruction/interpretation system. The motion segmentation system detects and tracks all moving objects, and removes the corresponding features from the original 2D feature data set. The remaining features are passed on to a 3D reconstruction system. The resulting list of 3D feature data, which hopefully contains useful information about the static part of the world, is then interpreted to give some high level understanding of the environment.

3D Reconstruction of Cerebral Vessels and Pathologies from a Few Biplane Digital Angiographies

Abstract: 3D reconstructions of cerebral vessels and Arteriovenous Malformations from six digital angiography sequences are compared for three iterative methods, one of them being new The interest of Multiplicative ART is underlined, and the effects of a criterion minimizing the total density are discussed

Curved surface reconstruction using monocular vision

Abstract: In monocular vision, a priori knowledge is necessary to perform 3D reconstruction. This paper describes how to evaluate two out of six external parameters of a camera in order to project an image on a curved surface (generalized cylinder). The final aim consists of reconstructing the model of the surface. Afterwards, with this model we can derive a flat representation of the scene without any distortions due to the projective geometry. In this work based on one projected view of the scene, we develop two methods to detect the projection of the revolution axis of the curved surface. With this axis, we can then extract the external parameters of a camera. The first one is based on the derivation of a polynomial function and the second one is based on the detection of the common normal between curves.

Visualization in biomedical computing : 4th International Conference, VBC '96, Hamburg, Germany, September 22-25, 1996 : proceedings

Abstract: GI tract unraveling in volumetric CT.- Segmentation of the visible human for high quality volume based visualization.- Illustrating anatomic models - A semi-interactive approach.- A fast rendering method using the tree structure of objects in virtualized bronchus endoscope system.- An optimal ray traversal scheme for visualizing colossal medical volumes.- A direct multi-volume rendering method. Application to visual assessment of 3-D image registration algorithms.- Visualization of tissue elasticity by magnetic resonance elastography.- Accurate vessel depiction with phase gradient algorithm in MR angiography.- Visualization and labeling of the Visible Human(TM) dataset: Challenges & resolves.- Rebuilding the visible man.- Image quality improvements in volume rendering.- Real-time merging of visible surfaces for display and segmentation.- Cortical peeling: CSF/grey/white matter boundaries visualized by nesting isosurfaces.- Enhancement of MR images.- Radiometric correction of color cryosection images for three-dimensional segmentation of fine structures.- 3D reconstruction of cerebral vessels and pathologies from a few biplane digital angiographies.- Robust 3-D reconstruction and analysis of microstructures from serial histologic sections, with emphasis on microvessels in prostate cancer.- An extensible MRI simulator for post-processing evaluation.- Compensation of spatial inhomogeneity in MRI based on a parametric bias estimate.- Analysis of coupled multi-image information in microscopy.- Bias field correction of breast MR images.- Region-growing based feature extraction algorithm for tree-like objects.- Scale-space boundary evolution initialized by cores.- Fast, accurate, and reproducible live-wire boundary extraction.- Automatic segmentation of cell nuclei from confocal laser scanning microscopy images.- Evaluation of segmentation algorithms for intravascular ultrasound images.- Fast algorithms for fitting multiresolution hybrid shape models to brain MRI.- Computer aided screening system for lung cancer based on helical CT images.- Unsupervised regularized classification of multi-spectral MRI.- Using an entropy similarity measure to enhance the quality of DSA images with an algorithm based on template matching.- Automatic segmentation of the brain in MRI.- Image analysis using modified self-organizing maps: Automated delineation of the left ventricular cavity boundary in serial echocardiograms.- Analyzing and predicting images through a neural network approach.- An artificial intelligence approach for automatic interpretation of maxillofacial CT images.- Fast Fluid Registration of medical images.- A robust point matching algorithm for autoradiograph alignment.- Registration updating using marker pins in a video based neurosurgical guidance system (VISLAN).- Point-based elastic registration of medical image data using approximating thin-plate splines.- Cortical constraints for non-linear cortical registration.- Image guided microscopic surgery system using mutual-information based registration.- Cross validation of three inter-patients matching methods.- A new approach to fast elastic alignment with applications to human brains.- Individualizing anatomical atlases of the head.- Mutual information for automated multimodal image warping.- Morphological analysis of brain structures using spatial normalization.- Three dimensional MR-based morphometric comparison of schizophrenic and normal cerebral ventricles.- Application of an automatically built 3D morphometric brain atlas: Study of cerebral ventricle shape.- Visualization and mapping of anatomic abnormalities using a probabilistic brain atlas based on random fluid transformations.- L-systems for three-dimensional anatomical modelling: Towards a virtual laboratory in anatomy.- Hierarchical data representation of lung to model morphology and function.- Visualizing group differences in outline shape: Methods from biometrics of landmark points.- Visualising cerebral asymmetry.- Assessing patterns and rates of brain atrophy by serial magnetic resonance imaging: A segmentation, registration, display and quantification procedure.- Characterisation and classification of brain tumours in three-dimensional MR image sequences.- Automatic quantification of multiple sclerosis lesion volume using stereotaxic space.- 3D skeleton for virtual colonoscopy.- Identifying hypometabolism in PET images of the brain: Application to epilepsy.- Multi-array EEG signals mapped with three dimensional images for clinical epilepsy studies.- Visualisation of pain by magnetoencephalography in humans.- Visualization of cat auditory cortical functional organization after electrical stimulation with a multichannel cochlear implant by means of optical imaging.- The brain bench: Virtual stereotaxis for rapid neurosurgery planning and training.- A flexible environment for image guided virtual surgery planning.- An advanced system for the simulation and planning of orthodontic treatments.- Assessment of several virtual endoscopy techniques using computed tomography and Perspective Volume Rendering.- Surgery Simulation using Fast Finite Elements.- Real time volumetric deformable models for surgery simulation.- Craniofacial surgery simulation.- Creation and validation of patient specific anatomical models for prostate surgery planning using virtual reality.- Use of shape for automated, optimized 3D radiosurgical treatment planning.- MRI guided intranasal flexible micro-endoscopy.- Computer-assisted insertion of pedicle screws.- PROBOT - A computer integrated prostatectomy system.- Towards performing ultrasound-guided needle biopsies from within a head-mounted display.- Automated multimodality registration using the full affine transformation: Application to MR and CT guided skull base surgery.

3D reconstruction using an uncalibrated stereo pair of encoded images

Abstract: The reconstruction of three dimensional (3-D) objects is used in CAD/CAM, robotics, remote sensing, etc. The models (images) can be either directly acquired by using special devices such as range finders, CTR scanners, etc. or they can be recovered from a series of 2-D images of the object. The authors report on a new method for reconstructing 3-D objects from two 2-D images using CCD cameras. This method uses a pseudo-random encoded grid projected on the object. The grid nodes are used in matching left to right 2-D images. The set of matched points are further used to calculate the disparity of each point of the object surface. Experimental examples illustrate the performance of this simple and elegant technique.

Cavity surface measuring system using stereo reconstruction

Abstract: 3D reconstruction of highly textured surfaces on unvegetated terrain is of major interest for stereo vision based mapping applications. We describe a prototype system for automatic modeling of such scenes. It is based on two frame CCD cameras, which are tightly attached to each other ensuring constant relative orientation. One camera is used to acquire known reference points to get the exterior orientation of the system, the other records the surface images. The system is portable to keep image acquisition as short as possible. Automatic calibration using the images acquired by the calibration camera permits the computation of exterior orientation parameters of the surface camera via a transformation matrix. A robust matching method providing dense disparities together with a flexible reconstruction algorithm renders an accurate grid of 3D points on arbitrarily shaped surfaces. The results of several stereo reconstructions are merged. Projection onto the global shape allows easy evaluation of volumes, and thematic mapping with respect to the desired surface geometry in construction processes. We report on accuracy and emphasize on the practical usage. It is shown that the prototpye system is able to generate a proper data set of surface descriptions that is accurate and dense enough to serve as documentation, planning and accounting basis.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.