Auto-calibration by direct observation of objects
Abstract
This paper presents techniques for using objects in a scene to define the reference frames for 3D reconstruction. We first present a simple technique to calibrate an orthographic projection from four non-coplanar reference points. We then show how the observation of two additional known scene points can provide the complete perspective projection. When used with a known object, this technique permits calibration of the full projective transformation matrix without matrix inversion. For an arbitrary non-coplanar set of four (or six) points, this calibration provides an affine basis for the reconstruction of local scene structure. When the four points define three orthogonal vectors, the basis is orthogonal, with a metric defined by the lengths of the three vectors. We demonstrate this technique for the case of a cube. We present results in which five and a half points on the cube are sufficient to compute the projective transformation for an orthogonal basis by direct observation (without matrix inversion). We then present experiments with three techniques for reducing the imprecision due to errors in the positions of the reference point because of pixel quantization and noise. We provide experimental measurements of the stability of the stereo reconstruction as a function of the error in the observed pixel position of the reference points used for calibration. A major problem in active 3D vision is updating a calibration matrix as the camera focus, aperture, zoom or vergence angle is changed. We present a technique for correcting the projective transformation matrix by tracking reference points. Our experiments show that correcting for a change of focus can be corrected by a affine transform, obtained by tracking three points. A camera rotation near the principal point, such as with stereo vergence, is slightly more precise with a projective correction matrix obtained by tracking at least four points. We then show how stereo reconstruction permits us to ‘hop’ the reference frame from the known calibration object to a new reference frame defined by reconstruction of four non-coplanar points. Any set of four non-coplanar points in the scene may define such a reference frame. We also show how to keep the reference frame locked onto a set of four points as the stereo head is translated or rotated. These techniques make it possible to reconstruct the shape of an object in its intrinsic coordinates without having to match new observations to a partially reconstructed description.
References (9)
- R.Y Tsai
A versatile camera calibration technique for high accuracy 3D machine vision metrology using off the shelf TV cameras and lenses
IEEE J. Robotics & Automat.
(August 1987) - O.D Faugeras et al.
The calibration problem for stereo
Comput. Vision & Patt. Recogn.
(June 1986) - P Puget et al.
Calibrating a mobile camera
Image & Vision Comput.
(November 1990) - J.L Crowley
Towards continuously operating integrated vision systems for robotics applications
Cited by (14)
The effect of noise on camera calibration parameters
2001, Graphical ModelsCamera calibration is the estimation of parameters (both intrinsic and extrinsic) associated with a camera being used for imaging. Given the world coordinates of a number of precisely placed points in a 3D space, camera calibration requires the measurement of the 2D projection of those scene points on the image plane. While the coordinates of the points in space can be known precisely, the image coordinates that are determined from the digital image are often inaccurate and hence noisy. In this paper, we look at the statistics of the behavior of the camera calibration parameters, which are important for stereo matching, when the image plane measurements are corrupted by noise. We derive analytically the behavior of the camera calibration matrix under noisy conditions and further show that the elements of the camera calibration matrix have a Gaussian distribution if the noise introduced into the measurement system is Gaussian. Under certain approximations we derive relationships between the camera calibration parameters and the noisy camera calibration matrix and compare it with Monte Carlo simulations.
A new approach to high precision 3-D measuring system
1999, Image and Vision ComputingA 3-D measuring system is important in many industrial applications like object modeling, medical diagnosis, CAD/CAM, multi-media, and virtual reality systems. However, it takes a lot of effort and cost to develop a high precision 3-D measuring system. A novel approach is introduced in this paper, with a view to develop a 3-D measuring system to meet the requirements of high performance through low cost equipment. A laser projector with controllable light intensity and phase is designed and used as the active light source. When the light is projected on an object, the varying depths of the object surface cause phase variations in the active light projected. These phase changes are used to find out the surface coordinates of the test object using simple Digital Signal Processing techniques. The advantages of the proposed method include the following: (1) as the measuring speed is fast, our method is suitable for on-line 3-D applications, e.g. the monitoring of human body or on-line detection; (2) only a single image is required to obtain the necessary phases, so the vibration errors can be avoided; (3) each CCD element is taken as a sample point, thus a high resolution result can be obtained; and (4) the measuring result is accurate. The experimental results reveal the superiority of the proposed method.
Fast relative depth computation for an active stereo vision system
1999, Real-Time ImagingReal-time computation of exact depth is not feasible in an active vision setup. Instead, reliable relative depth information which can be rapidly computed is preferred. In this paper, a stereo cue for computing relative depth obtained from an active stereo vision system is proposed. The proposed stereo cue can be computed purely from the coordinates of points in the stereo pair. The computational cost required is very low. No camera calibration or prior knowledge of the parameters of the stereo vision system is required. We show that the relationship between the relative depth cue and the actual depth in the three-dimensional (3D) space is monotonic. Such a relation is maintained even when the focal length and the vergence angle are changed, so long as the focal lengths of the two cameras are similar. Therefore, real-time implementation in an active vision setup can be realized. Stability analysis shows that the proposed method will be stable in practical situations, unless the stereo camera diverges. Experimental results are presented to highlight the properties and advantages of the proposed method.
Chapter 14 General theory of geometric estimation
1996, Machine Intelligence and Pattern RecognitionThe chapter presents a rigorous mathematical foundation for the theory of geometric correction and the theory of parametric fitting. For both, the problem is stated in general terms without assuming Gaussian noise: the role of the covariance matrix for a Gaussian distribution is played by the Fisher information matrix. A lower bound is derived that corresponds to the Cramer–Rao lower bound in traditional statistics, on the covariance matrix of the unbiased estimator of the parameter. Then, the maximum likelihood estimator is proved to attain it in the first order if the problem belongs to the exponential family. The maximum likelihood estimation process is expressed in a computationally convenient form, where the rank-constrained generalized inverse is used to discuss the ill-posedness of the problem and the numerical instability of the solution. The statistical problem closely related to the parametric fitting problem is what is known as the Neyman-Scott problem: observing multiple data, each having a distribution characterized by a common parameter, called the “structure parameter” or the “parameter of interest,” and a distinct parameter, called the “nuisance parameter,”) that depends on each observation, one has to estimate the structural parameter without knowing the nuisance parameters.
Image Analysis and Computer Vision: 1993
1994, Computer Vision and Image UnderstandingThis paper presents a bibliography of nearly 1300 references related to computer vision and image analysis, arranged by subject matter. The topics covered include computational techniques; feature detection and segmentation; image analysis; two-dimensional shape; pattern; color and texture; matching and stereo; three-dimensional recovery and analysis; three-dimensional shape; and motion. A few references are also given on related topics, such as geometry, graphics, coding and processing, sensors and optical processing, visual perception, neural nets, pattern recognition, and artificial intelligence, as well as on applications.
Auto-calibration of visual sensor parameters on a robotic head
1994, Image and Vision ComputingWe propose a new method of auto-calibration for a visual sensor mounted on a robotic head. The method is based on the tracking of stationary targets, while the robotic system, performs a specific controlled displacement, namely a fixed axis rotation. We derive here the equations related to this paradigm, and demonstrate that we can calibrate intrinsic and extrinsic parameters using this method. Experimental data is provided.