scispace - formally typeset
Search or ask a question

Showing papers on "Camera auto-calibration published in 1999"


Proceedings ArticleDOI
Zhengyou Zhang1
01 Sep 1999
TL;DR: Compared with classical techniques which use expensive equipment, such as two or three orthogonal planes, the proposed technique is easy to use and flexible, and advances 3D computer vision one step from laboratory environments to real-world use.
Abstract: Proposes a flexible new technique to easily calibrate a camera. It only requires the camera to observe a planar pattern shown at a few (at least two) different orientations. Either the camera or the planar pattern can be freely moved. The motion need not be known. Radial lens distortion is modeled. The proposed procedure consists of a closed-form solution followed by a nonlinear refinement based on the maximum likelihood criterion. Both computer simulation and real data have been used to test the proposed technique, and very good results have been obtained. Compared with classical techniques which use expensive equipment, such as two or three orthogonal planes, the proposed technique is easy to use and flexible. It advances 3D computer vision one step from laboratory environments to real-world use. The corresponding software is available from the author's Web page ( ).

2,661 citations


Journal ArticleDOI
TL;DR: A theoretical proof is given which shows that the absence of skew in the image plane is sufficient to allow for self-calibration and a method to detect critical motion sequences is proposed.
Abstract: In this paper the theoretical and practical feasibility of self-calibration in the presence of varying intrinsic camera parameters is under investigation. The paper‘s main contribution is to propose a self-calibration method which efficiently deals with all kinds of constraints on the intrinsic camera parameters. Within this framework a practical method is proposed which can retrieve metric reconstruction from image sequences obtained with uncalibrated zooming/focusing cameras. The feasibility of the approach is illustrated on real and synthetic examples. Besides this a theoretical proof is given which shows that the absence of skew in the image plane is sufficient to allow for self-calibration. A counting argument is developed which—depending on the set of constraints—gives the minimum sequence length for self-calibration and a method to detect critical motion sequences is proposed.

829 citations


Journal ArticleDOI
TL;DR: This paper demonstrates the feasibility of an end-to-end person tracking system using a unique combination of motion analysis on 3D geometry in different camera coordinates and other existing techniques in motion detection, segmentation, and pattern recognition.
Abstract: This paper presents a comprehensive framework for tracking coarse human models from sequences of synchronized monocular grayscale images in multiple camera coordinates. It demonstrates the feasibility of an end-to-end person tracking system using a unique combination of motion analysis on 3D geometry in different camera coordinates and other existing techniques in motion detection, segmentation, and pattern recognition. The system starts with tracking from a single camera view. When the system predicts that the active camera will no longer have a good view of the subject of interest, tracking will be switched to another camera which provides a better view and requires the least switching to continue tracking. The nonrigidity of the human body is addressed by matching points of the middle line of the human image, spatially and temporally, using Bayesian classification schemes. Multivariate normal distributions are employed to model class-conditional densities of the features for tracking, such as location, intensity, and geometric features. Limited degrees of occlusion are tolerated within the system. Experimental results using a prototype system are presented and the performance of the algorithm is evaluated to demonstrate its feasibility for real time applications.

356 citations


Proceedings ArticleDOI
16 Sep 1999
TL;DR: A simple, geometrically intuitive method which exploits the strong rigidity constraints of paral-lelism and orthogonality present in indoor and outdoor architectural scenes to recover the projection matrices for each viewpoint is proposed.
Abstract: We address the problem of recovering 3D models from uncalibrated images of architectural scenes. We propose a simple, geometrically intuitive method which exploits the strong rigidity constraints of paral-lelism and orthogonality present in indoor and outdoor architectural scenes. We present a n o vel algorithm that uses these simple constraints to recover the projection matrices for each viewpoint and relate our method to the algorithm of Caprile and Torre 2]. The projection matrices are used to recover partial 3D models of the scene and these can be used to visualise new viewpoints. Our approach d o e s not need any a priori information about the cameras being used. A w orking system called PhotoBuilder has been designed and implemented to allow a user to interactively build a VRML model of a building from uncalibrated images from arbitrary viewpoints 3, 4 ].

249 citations


Patent
09 Aug 1999
TL;DR: In this paper, a camera array captures plural component images which are combined into a single scene from which "panning" and "zooming" within the scene are performed, in one embodiment, each camera of the array is a fixed digital camera.
Abstract: A camera array captures plural component images which are combined into a single scene from which “panning” and “zooming” within the scene are performed. In one embodiment, each camera of the array is a fixed digital camera. The images from each camera are warped and blended such that the combined image is seamless with respect to each of the component images. Warping of the digital images is performed via pre-calculated non-dynamic equations that are calculated based on a registration of the camera array. The process of registering each camera in the arrays is performed either manually, by selecting corresponding points or sets of points in two or more images, or automatically, by presenting a source object (laser light source, for example) into a scene being captured by the camera array and registering positions of the source object as it appears in each of the images. The warping equations are calculated based on the registration data and each scene captured by the camera array is warped and combined using the same equations determined therefrom. A scene captured by the camera array is zoomed, or selectively steered to an area of interest. This zooming- or steering, being done in the digital domain is performed nearly instantaneously when compared to cameras with mechanical zoom and steering functions.

230 citations


Journal ArticleDOI
TL;DR: A stratified approach is proposed, which goes from projective through affine to metric, which allows retrieval of the affine calibration for constant intrinsic parameters and is also suited for use in conjunction with scene knowledge.
Abstract: In computer vision and especially for 3D reconstruction, one of the key issues is the retrieval of the calibration parameters of the camera. These are needed to obtain metric information about the scene from the camera. Often these parameters are obtained through cumbersome calibration procedures. There is a way to avoid explicit calibration of the camera. Self-calibration is based on finding the set of calibration parameters which satisfy some constraints (e.g., constant calibration parameters). Several techniques have been proposed but it often proved difficult to reach a metric calibration at once. Therefore, in the paper, a stratified approach is proposed, which goes from projective through affine to metric. The key concept to achieve this is the modulus constraint. It allows retrieval of the affine calibration for constant intrinsic parameters. It is also suited for use in conjunction with scene knowledge. In addition, if the affine calibration is known, it can also be used to cope with a changing focal length.

179 citations


Patent
17 Mar 1999
TL;DR: In this paper, a camera system for generating panoramic images is described, which uses sensors to determine the orientation of the camera as each image is recorded in memory, and a processor reconstructs a single panorama image from the recorded images using the recorded orientation information.
Abstract: A camera system for generating panoramic images is described. The camera uses sensors to determine the orientation of the camera as each image is recorded. The orientation information is recorded in memory. A processor reconstructs a single panoramic image from the recorded images using the recorded orientation information.

164 citations


Patent
29 Sep 1999
TL;DR: In this paper, a camera can be used to record multiple low-resolution images of an object by shifting a camera lens relative to an image sensor of the camera, each camera image represents a portion of the object.
Abstract: A camera can be used to record multiple low resolution images of an object by shifting a camera lens relative to an image sensor of the camera. Each camera image recorded represents a portion of the object. A composite high resolution image of the object can be obtained by patching together the camera images by using various well known mosaicing, tiling, and/or stitching algorithms.

155 citations


Proceedings ArticleDOI
24 Oct 1999
TL;DR: A new "CCD Fingerprint" method to identify a camcorder from videotaped images, which enables individual camera identification, which does not mean camera-model identification.
Abstract: We propose a new "CCD Fingerprint" method to identify a camcorder from videotaped images. This method enables individual camera identification, which does not mean camera-model identification. The clue to the individual camera identification is obtained from nonuniformity of dark currents on CCD chips, which causes a fixed pattern noise. The pattern is quite unique and intrinsic to an individual camera. Moreover, such a pattern might be recorded on tapes. Using these features, it is expected that the nonuniformity of dark current makes it possible to specify only one camera among same model cameras. In order to observe the characteristic noise patterns, we performed experiments with 9 camcorders in 4 types by recording a blank image on videotapes. The results showed that the fixed patterns were observed in most cases by accumulating 100 frames. Furthermore, it was confirmed that each camera had its own unique pattern.

136 citations


Proceedings ArticleDOI
23 Jun 1999
TL;DR: A linear self-calibration method is given for computing the calibration of a stationary but rotating camera, which is extremely rapid, and avoids the convergence problems characteristic of iterative algorithms.
Abstract: A linear self-calibration method is given for computing the calibration of a stationary but rotating camera. The internal parameters of the camera are allowed to vary from image to image, allowing for zooming (change of focal length) and possible variation of the principal point of the camera. In order for calibration to be possible some constraints must be placed on the calibration of each image. The method works under the minimal assumption of zero-skew (rectangular pixels), or the more restrictive but reasonable conditions of square pixels, known pixel aspect ratio, and known principal point. Being linear the algorithm is extremely rapid, and avoids the convergence problems characteristic of iterative algorithms.

136 citations


Proceedings ArticleDOI
01 Jan 1999
TL;DR: Imposing chirality constraints to limit the search for the plane at infinity to a 3-dimensional cubic region of parameter space is imposed and it is shown that this dense search allows one to avoid areas of local minima effectively and find global minima of the cost function.
Abstract: This paper considers the problem of self-calibration of a camera from an image sequence in the case where the camera's internal parameters (most notably focal length) may change The problem of camera self-calibration from a sequence of images has proven to be a difficult one in practice, due to the need ultimately to resort to non-linear methods, which have often proven to be unreliable In a stratified approach to self-calibration, a projective reconstruction is obtained first and this is successively refined first to an affine and then to a Euclidean (or metric) reconstruction It has been observed that the difficult step is to obtain the affine reconstruction, or equivalently to locate the plane at infinity in the projective coordinate frame The problem is inherently non-linear and requires iterative methods that risk not finding the optimal solution The present paper overcomes this difficulty by imposing chirality constraints to limit the search for the plane at infinity to a 3-dimensional cubic region of parameter space It is then possible to carry out a dense search over this cube in reasonable time For each hypothesised placement of the plane at infinity, the calibration problem is reduced to one of calibration of a nontranslating camera, for which fast non-iterative algorithms exist A cost function based on the result of the trial calibration is used to determine the best placement of the plane at infinity Because of the simplicity of each trial, speeds of over 10,000 trials per second are achieved on a 256 MHz processor It is shown that this dense search allows one to avoid areas of local minima effectively and find global minima of the cost function

Patent
15 Dec 1999
TL;DR: In this article, a method is presented in which three-dimensional models are created from a number of 2D images using a single camera without prior knowledge of the position or orientation of the camera, its focal length, etc.
Abstract: A method is provided in which three-dimensional models are created from a number of two-dimensional images, e.g., a video stream of two-dimensional images or still images, using a single camera without prior knowledge of the position or orientation of the camera, its focal length, etc. The method derives a relative value related to the position or orientation of the camera for each two-dimensional image and then, based on the derived position or orientation, uses 'ray casting' to develop the three-dimensional model based on intersecting rays through common features of sequential two-dimensional images.

Patent
15 Apr 1999
TL;DR: In this paper, a system for automatically monitoring an area includes a camera unit having a video camera and an image processing section, which detects and tracks change regions in the subsequent images.
Abstract: A system ( 10 ) for automatically monitoring an area includes a camera unit ( 12 ) having therein a video camera ( 23 ) and an image processing section ( 27 ). The image processing section saves a reference image from the video camera, compares subsequent images to the reference image, and detects and tracks change regions in the subsequent images. For each change region, the image processing section saves a path of movement of the change region, and a selected image of the change region. Selection is carried out so as to optimize the selected image, for example so that a detected person is facing and close to the video camera. The camera unit is network-ready ( 14 ), so that a remote workstation ( 13 ) can access the images and other information saved in the camera unit.

Proceedings ArticleDOI
05 Oct 1999
TL;DR: An automatic camera placement method for generating image-based models from scenes with known geometry that first approximately determines the set of surfaces visible from a given viewing area and then selects a small set of appropriate camera positions to sample the scene from.
Abstract: We present an automatic camera placement method for generating image-based models from scenes with known geometry. Our method first approximately determines the set of surfaces visible from a given viewing area and then selects a small set of appropriate camera positions to sample the scene from. We define a quality measure for a surface as seen, or covered, from the given viewing area. Along with each camera position, we store the set of surfaces which are best covered by this camera. Next, one reference view is generated from each reference view that do not belong to the selected set of polygons are masked out. The image-based model generated by our method, covers every visible surface only once, associating it with a camera position from which it is covered with quality that exceeds a user-specified quality threshold. The result is a compact non-redundant image-based model with controlled quality. The problem of covering every visible surface with a minimum number of cameras (guards) can be regarded as an extension to the well-known Art Gallery Problem. However, since the 3D polygonal model is textured, the camera-polygon visibility relation is not binary; instead, it has a weight-the quality of the polygon's coverage.

Patent
Zhengyou Zhang1
30 Apr 1999
TL;DR: In this paper, the intrinsic and extrinsic parameters of a camera mounted on a planar surface are estimated by analyzing the homographies associated with each image, and the intrinsic parameters for each image are computed from both intrinsic parameters and homographies.
Abstract: A digital camera is calibrated by establishing the coordinates of at least four feature points of a pattern mounted on a planar surface. At least two, and preferably three or more, images of the planar pattern are captured at different, non-parallel orientations using the digital camera. The image coordinates of the pattern's feature points are then identified in the captured images. A closed form solution can be employed to derive all the intrinsic and extrinsic parameters needed to provide the camera calibration. Essentially, the known pattern coordinates and corresponding image coordinates are used to compute a homography for each image. Then, a process is employed that estimates the intrinsic camera parameters by analyzing the homographies associated with each image. Finally, the extrinsic parameters for each image are computed from the intrinsic parameters and the homographies. However, the images can be effected by various noise sources which could affect the accuracy of the closed form solution process. If higher accuracy is called for, a maximum likelihood inference process can be employed to either provide a more accurate first estimate, or to refine the estimates derived from the closed form solution. If radial distortion caused by the lens of the camera is also a concern, the camera parameters can be further refined by taking into account this distortion.

Proceedings ArticleDOI
01 Jan 1999
TL;DR: This work extends the sequential camera tracking of an existing structure-from-motion approach to the calibration of a mesh of viewpoints, and proposes using an uncalibrated hand-held camera only to do this calibration.
Abstract: We focus on the calibration of very long image sequences from a hand-held camera that samples the viewing sphere of a scene. View sphere sampling is important for plenoptic (image-based) modeling that captures the appearance of a scene by storing images from all possible directions. The plenoptic approach is appealing since it allows, in principle, fast scene rendering of scenes with complex geometry and surface reflections, without the need for an explicit geometrical scene model. However the acquired images have to be calibrated, and current approaches mostly use pre-calibrated acquisition systems. This limits the generality of the approach. We propose using an uncalibrated hand-held camera only. The image sequence is acquired by simply waving the camera around the scene objects, creating a zigzag scan path over the viewing sphere. We extend the sequential camera tracking of an existing structure-from-motion approach to the calibration of a mesh of viewpoints. Novel views are generated by piecewise mapping and interpolating the new image from the nearest viewpoints according to the viewpoint mesh. Local depth map estimates enhance the rendering process. Extensive experiments with ground truth data and hand-held sequences confirm the performance of our approach.

Patent
11 May 1999
TL;DR: In this paper, the camera is moved in accordance with the moving angle to a position where the operable object and the other object existing in a photographed three-dimensional space are displayed on a display.
Abstract: A three-dimension image processing apparatus includes a CPU. When the CPU detects by collision determination that another object, e.g., a wall, is existent between an operable object and a camera, it calculates such a moving angle of the camera that an eye of the camera to the operable object is not obstructed by the other object. The camera is moved in accordance with the moving angle to a position where the operable object and the other object existing in a photographed three-dimensional space are displayed on a display.

Patent
04 Jan 1999
TL;DR: In this paper, an endoscopic camera system includes a scope for insertion into the body, a camera head coupled to the scope, and a camera control unit coupled with the camera head.
Abstract: An endoscopic camera system includes zooming capability to non-mechanically enlarge images of internal features of a body during an endoscopic procedure. The system includes a scope for insertion into the body, a camera head coupled to the scope, and a camera control unit coupled to the camera head. Video image data of internal features of the body are acquired through the scope by the camera head and transmitted to the camera control unit, which generates video images for display on a monitor. The camera control unit includes zoom circuitry for digitally zooming the video images in real-time. Operation of the zoom circuitry may be controlled by manual input or by voice commands. Voice commands may be used to set and recall preset zoom views.

Proceedings ArticleDOI
23 Jun 1999
TL;DR: A parametric camera model and calibration procedures are developed for an outdoor active camera system with pan, tilt and zoom control, and no special laboratory setups are required.
Abstract: A parametric camera model and calibration procedures are developed for an outdoor active camera system with pan, tilt and zoom control. Unlike traditional methods, active camera motion plays a key role in the calibration process, and no special laboratory setups are required. Intrinsic parameters are estimated automatically by fitting parametric models to the optic flow induced by rotating and zooming. No knowledge of 3D scene structure is needed. Extrinsic parameters are calculated by actively rotating the camera to sight a sparse set of surveyed landmarks over a virtual hemispherical field of view, yielding a well-conditioned pose estimation problem.

Patent
Bradford H. Needham1
21 Sep 1999
TL;DR: In this article, a motion detection algorithm captures a current image frame when a pixel comparison between successive image frames exceeds a predetermined threshold, i.e., the pixel comparison is larger than a certain threshold.
Abstract: A camera system that includes a video camera and a processor, which periodically uploads images captured by the video camera to a web server in accordance with one of a plurality of motion detection algorithms. A first motion detection algorithm captures a current image frame when a pixel comparison between successive image frames exceeds a predetermined threshold.

Journal ArticleDOI
TL;DR: A set of multichannel camera systems and algorithms is described for recovering both the surface spectral-reflectance function and the illuminant spectral-power distribution from the data of spectral imaging.
Abstract: A set of multi-channel camera systems and algorithms is described for recovering both the surface spectral- reflectance function and the illuminant spectral-power distribution from the data of spectral imaging. We show the camera system with six spectral channels of fixed wavelength bands. This system is created by using a monochrome CCD camera, six different color filters, and a personal computer. The dynamic range of the camera is extended for sensing the high intensity level of highlights. We suppose in a scene that the object surface of an inhomogeneous dielectric material is described by the dichromatic reflection model. The process for estimating the spectral information is composed of several steps of (1) the finite- dimensional linear model representation of wavelength functions, (2) illuminant estimation, (3) data normalization and image segmentation, (4) reflectance estimation. The reliability of the camera system and the algorithms is demonstrated in an experiment. Finally a new type of system using liquid crystal filters is briefly introduced.

Proceedings ArticleDOI
04 Oct 1999
TL;DR: An approach is described which achieves this goal by combining state-of-the-art algorithms for uncalibrated projective reconstruction, self-calibration and dense correspondence matching.
Abstract: Modeling of 3D objects from image sequences is a challenging problem and has been a research topic for many years. Important theoretical and algorithmic results were achieved that allow to extract even complex 3D scene models from images. One recent effort has been to reduce the amount of calibration and to avoid restrictions on the camera motion. In this contribution an approach is described which achieves this goal by combining state-of-the-art algorithms for uncalibrated projective reconstruction, self-calibration and dense correspondence matching.

Proceedings ArticleDOI
01 Jan 1999
TL;DR: The neurocalibration technique is shown to solve four different types of calibration problems that are found in computer vision applications and can be extended to the more difficult problem of calibrating cameras with automated active lenses.
Abstract: Camera calibration is a primary crucial step in many computer vision tasks. We present a new neural approach for camera calibration. Unlike some existing neural approaches, our calibrating network can tell the perspective-projection-transformation matrix between the world 3D points and the corresponding 2D image pixels. Starting from random initial weights, the net can specify the camera model parameters satisfying the orthogonality constraints on the rotational transformation. The neurocalibration technique is shown to solve four different types of calibration problems that are found in computer vision applications. Moreover, it can be extended to the more difficult problem of calibrating cameras with automated active lenses. The validity and performance of our technique are tested with both synthetic data under different noise conditions and with real images. Experiments have shown the accuracy and the efficiency of our neurocalibration technique.

Journal ArticleDOI
TL;DR: A method for the automatic reconstruction of 3D objects from multiple camera views for 3D multimedia applications that allows simultaneous acquisition of object and pattern, and ensures that each view can be calibrated individually.

Patent
15 Dec 1999
TL;DR: In this article, a digital still camera receives and digitizes visible radiation and sound waves from a scene, and when an operator actuates a shutter release, the camera detects and evaluates digitized information from the scene in a continuing manner, until a point in time when information representative of a human facial characteristic satisfies a specified criteria set by the operator through use of switches ( 47, 48 ) and a display ( 51 ).
Abstract: A digital still camera ( 10 ) receives and digitizes visible radiation ( 17 ) and sound waves ( 32 ) from a scene ( 12 ). When an operator actuates a shutter release ( 46 ), the camera detects and evaluates digitized information ( 22, 38 ) from the scene in a continuing manner, until a point in time when information representative of a human facial characteristic satisfies a specified criteria set by the operator through use of switches ( 47, 48 ) and a display ( 51 ). The camera then records in a memory ( 59 ) a digital image of the scene, which corresponds to the point in time. The stored image is a high-quality image, which can avoid characteristics such as eyes that are closed or a mouth that is open.

Patent
29 Apr 1999
TL;DR: In this article, a system for automatically monitoring an area includes a camera unit (12) having therein a video camera and an image processing section (27), which detects and tracks change regions in subsequent images.
Abstract: A system (10) for automatically monitoring an area includes a camera unit (12) having therein a video camera (23) and an image processing section (27). The image processing section saves a reference image from the video camera, compares subsequent images to the reference image, and detects and tracks change regions in the subsequent images. For each change region, the image processing section saves a path of movement of the change region, and a selected image of the change region. Selection is carried out so as to optimize the selected image, for example so that a detected person is facing and close to the video camera. The camera unit is network-ready (14), so that a remote workstation (13) can access the images and other information saved in the camera unit.

Patent
13 May 1999
TL;DR: In this article, an image correction apparatus for a stereo camera which corrects optical positional deviations in images taken by the stereo camera is proposed. But the method is not suitable for stereo matching.
Abstract: An image correction apparatus for a stereo camera which corrects optical positional deviations in images taken by the stereo camera, comprising: means for performing interpolation on mutual referencing positions in stereo matching of a pair of images taken by the stereo camera by using pixel-by-pixel positional deviations that have been determined in advance and coordinate data around the referencing positions.

Patent
24 Aug 1999
TL;DR: In this paper, the angular velocity sensor takes a certain time to be stabilized after its operation is started, and when correction of camera-shake is required during the certain period, it is displayed that correction is not possible.
Abstract: A camera comprises an angular velocity sensor for detecting camera-shake. The angular velocity sensor takes a certain time to be stabilized after its operation is started. When correction of camera-shake is required during the certain period, it is displayed that correction of camera-shake is not possible. As a result, a photographer using the camera with the camera-shake detection sensor refrains from photographing when correction of camera-shake is not possible.

Journal ArticleDOI
TL;DR: A dynamic sensor-planning system that is capable of planning the locations and settings of vision sensors for use in an environment containing objects moving in known ways is presented.
Abstract: This paper presents a dynamic sensor-planning system that is capable of planning the locations and settings of vision sensors for use in an environment containing objects moving in known ways. The key component of this research is the computation of the camera position, orientation, and optical settings to be used over a time interval. A new algorithm is presented for viewpoint computation which ensures that the feature-detectability constraints of focus, resolution, field of view, and visibility are satisfied. A five-degree-of-freedom Cartesian robot carrying a CCD camera in a hand/eye configuration and surrounding the work cell of a Puma 560 robot was constructed for performing sensor-planning experiments. The results of these experiments, demonstrating the use of this system in a robot work cell, are presented.

Patent
05 Mar 1999
TL;DR: In this paper, an apparatus for controlling a camera to achieve sharp, clear pictures includes a sensor for determining the distance of the camera from a subject and a controller for calculating the camera's depth-of-field, determining whether the subject is within the depth of field, and generating various signals which are provided to control operation of camera or convey pertinent information to the photographer.
Abstract: An apparatus for controlling a camera to achieve sharp, clear pictures includes a sensor for determining the distance of the camera from a subject and a controller for calculating the camera's depth-of-field, determining whether the subject is within the depth-of-field, and generating various signals which are provided to control operation of the camera or convey pertinent information to the photographer.