Title: Important approach to 3D reconstruction of tridimensional objects based on multiple plan images
Authors: Boutaina Satouri; Abdellatif El Abderrahmani; Khalid Satori
Addresses: LIIAN, Department of Computer Sciences, Dhar-Mahraz Sciences School, University Sidi Mohammed Ben Abdellah, P.B. 1796 Atlas-Fez, Morocco ' Departement of Computer Sciences, Poly disciplinary School University, Abdelmalek Essaadi, Larache, Morocco ' LIIAN, Department of Computer Sciences, Dhar-Mahraz Sciences School, University Sidi Mohammed Ben Abdellah, P.B. 1796 Atlas-Fez, Morocco
Abstract: In the present paper, we will focus on a new approach for efficient and reliable tridimensional reconstruction of objects from flat images. Our approach allows the realisation of tridimensional reconstruction without passing through the calibration and self-calibration phase of the camera, but based on the estimation of the fundamental matrix and the homography at infinity to have the projective, affine and Euclidean projection. Our method is based firstly on a very important step in the 3D reconstruction that is the detection of interest points using the Harris detector to have a sufficient number of matches distributed on the images, these matches are used to estimate the 3D points, and secondly to estimate the projection matrices that are made from different existing relationships between the three types of tridimensional reconstruction (projective reconstruction, affine reconstruction, Euclidean reconstruction). Experimental results prove that this method is practical and gives satisfying results without going through the calibration step.
Keywords: projective reconstruction; affine reconstruction; Euclidean reconstruction; interests points; matching; fundamental matrix.
International Journal of Computer Applications in Technology, 2017 Vol.56 No.2, pp.152 - 162
Received: 31 Aug 2016
Accepted: 22 Nov 2016
Published online: 13 Oct 2017 *