Analysis of Three Dimensional Object Reconstruction Method for Underwater Images

Abstract The aim of this paper is to design and develop a three dimensional reconstruction system using a feature detection technique called as speeded up robust features SURF. First the feature points are detected using SURF and matched between pair of underwater stereo images. Further epipolar geometry is utilized to remove the unfavourable feature points and to derive geometrical relation. Linear triangulation is used to compute 3d points which are further dense matched and mapped with the appropriate texture to resemble the original object. The proposed design is tested on a set of underwater images using the information about the camera calibration parameters. The result of testing shows that the designed method can be utilized to reconstruct a partial 3D approximation of the underwater objects.