Résumé
This work presents the 3D reconstruction of the archeology objects in underwater environment. The videos and images are obtained from a calibrated camera system. The features of interest between image pairs are selected by the FAST detector and FLANN descriptor. Subsequently, the RANSAC method is applied to reject outlier points. The putative inliers are matched by triangulation to produce sparse point clouds in 3D space, using a pinhole camera model and Euclidean distance estimation. The 3D model textrue is achieved through a commond 3D Delaunay triangulation procedure. Moreover, evaluation of the quality of the underwater 3D reconstruction model to find its accuracy is done by measurement, comparing it with a 3D industrial scan model as well as a real object.