Correspondence problem

Abstract: The present invention provides a method for solving the problem of correspondence associated with stereo cameras. The method includes the steps of aligning a central horizontal epipolar line parallel to the scan line of the CCD; choosing at least one scan line of the CCD that adheres to epipolar geometry, conduction of a disparity estimation of the scan line chosen, self calibration of the camera, and rectification of the images. In an alternative embodiment of the present invention, a laser pointer is aimed at the epipolar plane positioned perpendicular to the base line to help aim the device at the image or measurement volume, and allow the camera to adjust itself.

Abstract: The notion of correspondence underlies many current theories of human and machine visual information processing. Algorithms for both the correspondence process and solutions to the correspondence problem have appeared regularly in the computer vision literature. Algorithms for stereopsis (Marr and Poggio, 1977; Barnard and Thompson, 1980; Mayhew and Frisby, 1980) and for tracking objects through time (Moravec, 1977; Ullman, 1979; Dreschler and Nagel, 1981; Webb, 1981; Jain and Sethi, 1984) have been presented which assume that token matching of separated or successive views is the underlying visual process. This paper will address the notion of token matching as a primitive operation in vision. We will argue that correspondence seems ill suited to the task of accounting for how an object is positioned in time or space, and that some other mechanism may provide a more apt account.

Abstract: Data association or the correspondence problem is often considered as one of the key challenges in every state estimation algorithm in robotics. This work introduces an efficient multi-dimensional assignment based data association algorithm for simultaneous localization and map building (SLAM) problem in mobile robot navigation. Data association in SLAM problem is compared with the data association in a multi-sensor multi-target tracking context and formulated as a 0-1 integer programming (IP) problem. A suboptimal dual frame assignment based data association scheme is thus formulated using a linear programming relaxation of the IP problem. Simulations were conducted to verify the superior nature of the new data association scheme over the conventional nearest neighbor data association algorithm in the presence of high clutter densities. Experimental results are also presented to verify the enhanced performance of the algorithm.

Abstract: Finding correspondences between image pairs is a fundamental task in computer vision. Herein, we focus on establishing matches between images of urban scenes which are typically composed of planar surface patches with highly repetitive structures. The latter property makes traditional point-based methods unreliable. The basic idea of our approach is to formulate the correspondence problem in terms of homography estimation between planar image regions: given a planar region in one image, we are simultaneously looking for its corresponding segmentation in the other image and the planar homography acting between the two regions. We will show, that due to the overlapping views the general 8 degree of freedom (DOF) of the homography mapping can be geometrically constrained to 3 DOF and the resulting segmentation/registration problem can be efficiently solved by finding the region's occurrence in the second image using pyramid representation and normalized mutual information as the intensity similarity measure. The method has been validated on a large database of building images taken by different mobile cameras and quantitative evaluation confirms robustness against intensity variations, occlusions or the presence of non-planar parts. We also show examples of 3D planar surface reconstruction as well as 2D mosaicking.