Multicamera rigs are used in a large number of 3D Vision applications, such as 3D modeling, motion capture or telepresence and a robust calibration is of utmost importance in order to achieve a high accuracy results. In many practical configurations the cameras in a rig are arranged in such a way, that they can observe each other, in other words a number of epipoles correspond to the real image points. In this paper we propose a solution for the automatic recovery of the external calibration of a multicamera system by enforcing only simple geometrical constraints, arising from the epipole visibility, without using any calibration object, such as checkerboards, laser pointers or similar. Additionally, we introduce an extension of the method that handles the case of epipoles being visible in the reflection of a planar mirror, which makes the algorithm suitable for the calibration of any multicamera system, irrespective of the number of cameras and their actual mutual visibility, and furthermore we remark that it requires only one or a few images per camera and therefore features a high speed and usability. We produce an evidence of the algorithm effectiveness by presenting a wide set of tests performed on synthetic as well as real datasets and we compare the results with those obtained using a traditional LED-based algorithm. The real datasets have been captured using a multicamera Virtual Reality (VR) rig and a spherical dome configuration for 3D reconstruction.

Camera calibration denotes the task of estimating the projective mapping between 3D world and the camera image plane. Most of the modern cameras have multiple image and video acquisition modes, which differ in image resolution, field of view or aspect ratio. Each mode should be treated as an independent device, and therefore independently calibrated. This straightforward solution implies the acquisition of the calibration dataset and execution of the calibration routine for each of the camera modes, a time-consuming and error-prone task, especially in a multi-camera scenario. In this paper we propose a multimode camera calibration approach that noticeably relieves this task. We show how the calibration model of one camera mode can be easily transferred to the other modes without the need of multiple calibrations. The evaluation test performed on off-the-shelf consumer cameras shows that the proposed method not only reduces the required amount of user interaction, but also allows for the calibration accuracy improvement.

Non-metric camera calibration is the art of modeling the geometry of the image formation process using only qualitative constraints rather than the metric knowledge of the scene. In this paper we present a new color-coded calibration pattern, specifically designed for non-metric calibration. It embeds two bundles of orthogonal lines in two different color channels granting a two major advantages. Each image containing such a pattern naturally carries all the geometrical constraints needed for the calibration process, while the complexity of the pattern detection is reduced by decoupling the recovery of the two line bundles in two independent processing phases. Furthermore, we present an analysis of the displacement between corresponding lines detected in the color channels and in the grayscale image and propose a simple technique, based on a revised version of the Inverse Compositional Algorithm (ICA), as a solution for this issue. The performance of the proposed pattern was evaluated using the datasets, captured with a wide angle lens camera. The results suggest, that the use of a color-coded calibration pattern not only significantly reduces the amount of required user interaction, but also helps to achieve a high accuracy camera calibration, providing an important contribution to the integration of a fully automatic camera calibration tool.

The paper presents the applicability of the Rodrigues Rotation Formula (RRF) in the context of Two-Views Geometry estimation. The Epipolar Constraint is usually formulated as the belonging of an image point to its corresponding Epipolar Line, instead using the RRF we will arrange the same constraint in terms of equivalence between 3D unit-norm vectors. This alternative formulation generates a geometrically meaningful error measure, which according to the presented results, provides a significant improvement of the estimation accuracy with respect to the widely used Sampson error, still maintaining a comparable computational cost.

The paper presents a novel area-based technique aimed to the accurate subpixel estimation of straight lines, whose central and innovative aspect is represented by the minimization of an error function comprising a conveniently weighted version of the image response to the Gauss-Laguerre Circular Harmonic filter (1,0), which presents a high selectivity on border features. The paper presents a wide set of results, obtained in different contexts like highly noisy images and fast visual tracking, which confirm the robustness of the proposed technique and suggest its applicability as basic block in any application which could potentially benefit from the availability of accurate localization of long straight segments or short edgelet features.

The present paper is devoted to the introduction of a novel technique to select keypoints from digital images and build representative and distinctive descriptors. The algorithm performs a multiresolution image analysis in the Laguerre Gauss transformed domain and collects in a local descriptor the transformed coefficients at multiple scales, of the keypoint’s representative pattern. The rotation invariance of the Circular Harmonics and the multiscale approach make the system more robust than other descriptors to match patterns related by affine transformations. 1

The determination of 3D structure from images is one of the most active field in computer vision. This paper gives a contribution presenting a symmetric stereo matching algorithm, based on a bidirectional dynamic programming scanline optimization. Consistent matches are detected by inspecting the images scanlines with a complex exponential weighting window and then analyzing matching score map into a bidirectional optimization framework. Results are presented running the algorithm on stereo pair freely available on the Web, in order to numerically validate the performance.