Showing papers on "Homography (computer vision) published in 2002"

Scalable alignment of large-format multi-projector displays using camera homography trees

Abstract: This paper presents a vision-based geometric alignment system for aligning the projectors in an arbitrarily large display wall. Existing algorithms typically rely on a single camera view and degrade in accuracy1 as the display resolution exceeds the camera resolution by several orders of magnitude. Naive approaches to integrating multiple zoomed camera views fail since small errors in aligning adjacent views propagate quickly over the display surface to create glaring discontinuities. Our algorithm builds and refines a camera homography2 tree to automatically register any number of uncalibrated camera images; the resulting system is both faster and significantly more accurate than competing approaches, reliably achieving alignment errors of 0.55 pixels on a 24-projector display in under 9 minutes. Detailed experiments compare our system to two recent display wall alignment algorithms, both on our 18 Megapixel display wall and in simulation. These results indicate that our approach achieves sub-pixel accuracy even on displays with hundreds of projectors.

Self-Calibration of Rotating and Zooming Cameras

Abstract: In this paper we describe the theory and practice of self-calibration of cameras which are fixed in location and may freely rotate while changing their internal parameters by zooming. The basis of our approach is to make use of the so-called infinite homography constraint which relates the unknown calibration matrices to the computed inter-image homographies. In order for the calibration to be possible some constraints must be placed on the internal parameters of the camera. We present various self-calibration methods. First an iterative non-linear method is described which is very versatile in terms of the constraints that may be imposed on the camera calibration: each of the camera parameters may be assumed to be known, constant throughout the sequence but unknown, or free to vary. Secondly, we describe a fast linear method which works under the minimal assumption of zero camera skew or the more restrictive conditions of square pixels (zero skew and known aspect ratio) or known principal point. We show experimental results on both synthetic and real image sequences (where ground truth data was available) to assess the accuracy and the stability of the algorithms and to compare the result of applying different constraints on the camera parameters. We also derive an optimal Maximum Likelihood estimator for the calibration and the motion parameters. Prior knowledge about the distribution of the estimated parameters (such as the location of the principal point) may also be incorporated via Maximum a Posteriori estimation. We then identify some near-ambiguities that arise under rotational motions showing that coupled changes of certain parameters are barely observable making them indistinguishable. Finally we study the negative effect of radial distortion in the self-calibration process and point out some possible solutions to it.

Surveillance system and methods regarding same

Abstract: A system and method for use in monitoring a search area includes positioning a plurality of imaging devices at one or more installation sites such that corresponding fields of view for the plurality of imaging devices cover a defined search area. For example, each field of view of each imaging device may include a field of view portion which overlaps with at least one other field of view of another imaging device, e.g., the field of view portion which overlaps may be greater than about 25 percent and less than about 85 percent of the field of view of the imaging device. Further, for example, images may be fused using one or more homography transformation matrices.

Augmented reality camera tracking with homographies

Abstract: To realistically integrate 3D graphics into an unprepared environment, camera position must be estimated by tracking natural image features. We apply our technique to cases where feature positions in adjacent frames of an image sequence are related by a homography, or projective transformation. We describe this transformation's computation and demonstrate several applications. First, we use an augmented notice board to explain how a homography, between two images of a planar scene, completely determines the relative camera positions. Second, we show that the homography can also recover pure camera rotations, and we use this to develop an outdoor AR tracking system. Third, we use the system to measure head rotation and form a simple low-cost virtual reality (VR) tracking solution.

Visual navigation using planar homographies

Abstract: We introduce three new results, which allow homographies of the ground plane to support visual navigation functions for mobile robots using uncalibrated cameras. Firstly, we illustrate how, for pure translation, a homography can be computed from just two pairs of corresponding corner features. Secondly, we show how, for pure translation, we can determine the height of corner features above the ground plane using the recovered homography and a construct based on the cross ratio. This allows us to detect points which can be driven over, as their height is measured to be close to zero, and points which are sufficiently high to drive under. Finally, we show how, in the case of general planar motion, homographies can be used to determine the rotation of the camera and robot.

Intermediate view generation of soccer scene from multiple videos

Abstract: This paper introduces a novel method for generating an intermediate view of soccer scenes taken by multiple video cameras. In the proposed method, a soccer scene is classified into dynamic regions, a field region, and a background region. Using an epipolar geometry in the first region and homography in the second, dense correspondence is obtained to interpolate views. For the third region, partial area images are extracted from the panoramic image compounded from the background of multiple views. Finally synthesizing them completes intermediate view images of the whole object. Applying this method to actual scenes of a soccer match captured at the stadium, we succeeded in generating natural intermediate view videos.

Visual servo control using homography estimation for the stabilization of an X4-flyer

Abstract: In this paper we study image based visual servo (IBVS) control of a four rotor vertical take-off and landing (VTOL) craft known as an X4-flyer. The approach taken is tailored to planar visual targets and is based on a reformulation and extension of 2 1/2 D image-based visual servo design to the control of dynamic under-actuated systems. Exponential stability of the closed loop system is proved and simulation results are presented.

Homography transfer from point matches

Abstract: An infinity homography for an image pair within an image sequence is transferred to other image pairs within the image sequence utilizing point matches for the subject image pairs. An image set including the image pair for which the infinity homography is known and a third image are selected. Then intermediate parameters for homography transfer for one image pair overlapping the known infinity homography image pair is computed from the known infinity homography and epipoles and fundamental matrices of the overlapping image pairs, to derive the infinity homography for the selected overlapping image pair. The process may then be repeated until infinity homographies for all image pairs of interest within the image sequence have been derived.

An empirical comparison of methods for image-based motion estimation

Abstract: This paper presents a comparison between methods that estimate motion of a camera from a sequence of video images. We implemented two methods: a homography based method that assumes planar environments; and shape-from-motion, a general method that can deal with a fully three dimensional world. Both methods were formulated in an iterative, online form to produce estimates of camera motion. We discuss a trade-off in accuracy and run time efficiency based on experimental results for these two general methods in relation to ground truth. We show how a variation of the homography method can produce accurate results in some cases when the environment is non-planar with low computational cost.

A switching approach to visual servo control

Abstract: In the recent past, many researchers have developed control algorithms for visual servo applications. In this paper we introduce a new switching approach, in which a high-level decision maker determines which of two low level visual servo controllers should be used at each control cycle. We introduce two new low-level controllers, one that relies on the homography between initial and goal images, and one that uses an affine transformation to approximate the motion between initial and goal camera configurations. Since an affine transformation can only approximate a restricted set of camera motions, this choice of low-level controllers illustrates the strength of the switching approach. We evaluated our approach with several simulations using three candidate switching rules. Although our results are preliminary, we believe that the proposed method is very promising for visual servo tasks in which there is a significant distance between the initial and goal configurations.

Multiview Constraints for Recognition of Planar Curves in Fourier Domain

Abstract: Multiview relations such as the Fundamental matrix and the Trilinear tensor provide scene-independent characterization of a combination of views in the form of algebraic constraints. In this paper, we present a number of multiview constraints for collections of primitives, such as a planar shape boundary. The rich domain of Fourier transforms helps us to combine the properties of the collection with the multiview situation. We derive a number of viewindependent algebraic constraints under the assumption of affine image-to-image homography. These constraints provide useful tools to match and recognize planar boundaries across multiple views without the knowledge of the camera parameters or pixel-to-pixel correspondence. We present the results of shape matching in a number of synthetic and real situations in this paper.

Linear determination of the infinite homography and camera self-calibration

Abstract: In this paper, a new constraint on the homography of the plane at infinity is introduced and a new linear camera calibration technique is proposed based on it. Compared with the related techniques in the literature, the main advantages of our new technique are tow-fold. Firstly, it is less stringent to hardware, for example, it does not require the camera to undertake orthogonal motions which are usually difficult to be done without special hardware support. In contrast, our technique requires only one translation, and two general motions of camera, which can be easily done, for example, with a hand-held camera. Secondly, in the determination of the homography of the plane at infinity, it relies neither on projective reconstruction nor on the homography of a space plane, it needs only some image correspondences and epipoles, which are basic requirements for any camera self-calibration technique. In addition, we prove that for a given set of camera motions such as {(R,t1),(R,t2)},if (t1,t2) are not linearly dependent, then the homography of the plane at infinity under this motion set can be linearly and uniquely determined. Simulations and experiments with real images validate our new method.

A semi-automatic camera calibration method for augmented reality

Abstract: Camera calibration is a fundamental problem in computer vision. In a camera calibration process, the correspondences between 3D feature points and their 2D projection points must be obtained. We developed a semi-automatic calibration method for our augmented reality system, which uses corner detection to extract pixel coordinates of the projection points and uses homography to build the correspondences. Experiments show that the method is very easy to use and has good practical accuracy.

Computer vision techniques for fire monitoring using aerial images

Abstract: The paper presents an approach to the use of aerial images for fire monitoring. It shows techniques to segment the fire on visual images and to geo-locate the fire front. Color processing is used for fire segmentation. To geo-locate the fire base, a planar surface is assumed. The homography between the image plane and the real plane can be computed from several points correspondences. This homography is later used to geo-locate the fire base. Several artificial landmarks are used to obtain an initial relation. The homography is updated as the camera moves by tracking several points over the sequence of images. The points are tracked by using a feature matching algorithm. The explained procedure has been applied to visual images of controlled field fires taken by a camera placed on a helicopter.

A novel approach for single view based plane metrology

Abstract: An approach is proposed for single view based plane metrology. The approach is based on a pair of vanishing points from two orthogonal sets of space parallel lines. Extensive experiments on simulated data as well as on real images showed that our new approach can achieve as good result as that of the homography based one which is widely used in the literature, but our new approach does not need any explicit specifications of space control points. Since in many real applications, particularly, in indoor environment, orthogonal lines are not rare, for example, a frame of window or a door, our new approach is of widely applicable.

Ground plane segmentation from multiple visual cues

Abstract: An approach which uses multiple sources of visual information (or visual cues) to identify and segment the ground plane in indoor mobile robot visual navigation applications is presented. Information from color, contours and corners and their motion are applied, in conjunction with planar homography relations, to identify the navigable area of the ground, which may be textured or non-textured. We have developed new algorithms for both the computation of the homography, in which a highly stable two point method for pure translation is proposed, and the region growing. Also, a new method for applying the homography to measure the height of a visual feature to the ground using an uncalibrated camera is also developed. Regions are segmented by color and also by their sizes and geometric relation and these region boundarys are extracted as contours. By controlled manoeuvres of a mobile robot, the methods of coplanar feature grouping developed in this paper are not only applicable to corner correspondences but also to contours. This leads to robust, accurate segmentation of the ground plane from the other image regions. Results are presented which show the validity of the approach.

Robust Image Matching under a Large Disparity

Abstract: We present a new method for detecting point matches between two images without using any combinatorial search. Our strategy is to impose various local and non-local constraints as "soft" constraints by introducing their "confidence" measures via "mean-field approximations". The computation is a cascade of evaluating the confidence values and sorting according to them. In the end, we impose the "hard" epipolar constraint by RANSAC. We also introduce a model selection procedure to test if the image mapping can be regarded as a homography. We demonstrate the effectiveness of our method by real image examples.

Self-Calibration of a Camera with a Non-Linear Model

Abstract: Camera calibration is an indispensable step to obtain 3D geometric information from 2D images. Self calibration of a camera with a linear model (hereafter called a linear camera) has become one of the major research directions in computer vision field. However, to our knowledge, there has been no reports in the literature on self calibrating a camera with a non linear model (hereafter called a non linear camera). Since a linear camera model cannot in general accurately represent the imaging geometry of a real camera, it is both necessary and important to explore ways of self calibrating a non linear camera. In this paper, a self calibration method of a non linear camera is proposed. The basic principle of the method is to model the non linear camera as a linear camera plus a non linear distortion part, then a set of non linear constraints on the non intrinsic parameters is derived by means of fundamental matrices or homographies associated with linear camera. Experiments with simulated data as well as with real images show that the proposed method is workable, and applicable in real applications.

An Uncalibrated Lightfield Acquisition System.

Abstract: Acquisition of image data for lightfield usually requires an expensive, complex and bulky setup. In this paper, we describe a simple method of acquiring the image data set. The method requires a normal handheld video camera, which is taken around the object to be rendered. We employ homography from the viewing/camera plane to the lightfield plane for obtaining the ray intersections with the lightfield planes. The computations involved are simple and make the method suitable for online lightfield acquisition.

Region segmentation and matching in stereo images

Abstract: We propose a new method to simultaneously achieve segmentation and dense matching in a pair of stereo images. In constrast to conventional methods that are based on similarity or correlation techniques, this method is based on geometry and uses correlations only on a limited number of key points. Stemming from the observation that our environment is abundant in planes, this method focuses on segmentation and matching of planes in an observed scene. Neither prior knowledge about the scene nor camera calibration are needed. Using two uncalibrated images as inputs, the method starts with a rough identification of a potential plane, defined by three points only. Based on these three points, a plane homography is then calculated and used for validation. Starting from a seed region defined by the original three points, the method grows the current region by successive move/confirmation steps until occlusions and/or surface discontinuity occur. In this case, the homography-based mapping of points between the two images will fail. This failure is detected by the correlation, used in the confirmation process. In particular this method grows a region even across different colors as long as the region is planar. Experiments on real images validated our method and showed its capability and performance.

Plane based relative structure recovery

Abstract: In this paper we provide a novel linear algorithm to estimate the direction of translation and the normal to the scene plane, given a stereo pair with known correspondences between the corners in the images. Planar parallax which arises naturally in planar scenes (and can also be defined with respect to virtual planes), is exploited to achieve this. Based on the parallax vectors and the homography induced by the scene plane we calculate the direction of translation and the normal to the scene plane. We also calculate the relative heights of points on the object with respect to the scene plane, given the height of the camera above the plane. The accuracy of the estimated normal with respect to noise in input data is also analyzed empirically.

Омофонический и омографический фонды современного русского языка. Часть 4: функциональный аспект

Abstract: Homophony and homography of Russian are in clause analyzed as the phenomenon with a wide range of speech functioning. In turn this range expands structural borders of the given phenomens. The wide interpretation of homophony and homography are based on a subjective nature of interpretation of the written down and readable texts during their perception (recognition) and understanding. For this reason Russian hornophony and homography represent convenient test range for research of various problems of perception (recognition), identification and understanding of the text. In clause the illustration of such research opportunities Russian homophony and homography are carried out.

Direct Method for Motion Estimation: An Alternative to Decomposition of Planar Transformation Matrices

Abstract: Recently, we have seen a proliferation in research addressing the motion estimation problem and its practical applications based on coplanar point configurations [10,14,12,9,6,7,13]. This paper introduces a new approach where following an estimation of the full projection matrix from non-coplanar points in one reference frame, the system provides better motion estimation results based on coplanar point configurations without estimating the camera intrinsic parameters. The new mathematical framework allows us to directly estimate the rigid transformation between a full projection matrix and a homography. Experimental results compare the accuracy of this and the homography decomposition approach, proposed in [14], in the context of an augmented reality application where three cameras are calibrated for real-time image augmentation.

Two-step nonlinear medical image registration based on image intensity

Abstract: A two-step non-linear medical image registration approach is proposed, based on the image intensity. In the first step, the global affme medical image registration is used to establish one-to-one mapping between the two images to be registered. After this first step, the images are registered up to small local elastic deformation. Then the mapped images are used as inputs in the second step, during which, the study image is modeled as elastic sheet by being divided into several sub-images. Moving the individual sub-image in the reference image, the local displacement vectors are found and the global elastic transformation is achieved by assimilating all of the local transformation into a continuous transformation. This algorithm has been tested by both simulated and clinical tomographic images.

Algorithm for viewpoint estimation and registration

Abstract: This paper presents a simple technique for estimating the space location from which a certain image has been taken. The basic assumption is that the scene portrayed in the image is planar. The method is based on the acquisition of a new set of images, closely resembling the given image. The location is recovered from the parameters describing the camera's pose during the acquisition of that among the new images showing the highest degree of correlation with the original image. An example of application of this technique is discussed in the paper.