Analysis and Development of Stereoscopic Vision System

Abstract

The focus of the thesis is the study of the geometry involved in stereoscopic vision system composed of two cameras and object with the aim of obtaining 3D geometric information of the object. This thesis deals with the study of camera calibration and the comprehension of the Epipolar geometry to calculate the depth of every point of the object in the scene.

There are mainly four parts for calculating the depth of every point of object. (1) Camera calibration which deals with the calculation of the intrinsic and extrinsic parameters of both cameras. (2) Point correspondence is used to find out pair of points which match between both images. This section also deals with the removal of unwanted and unmatched points. (3) Study of different methods of calculation of Fundamental matrix for obtaining epipolar constraint. The computation of the fundamental matrix of a stereoscopic vision system is surveyed with the aim of reducing correspondence problem between both images. (4) Depth perception can be calculated using the geometry between two camera and object in which focal length, distance between two camera centers, and disparity are main parameters.

Keywords: Camera calibration, Epipolar Geometry, Fundamental Matrix, Point Correspondence, Intrinsic and Extrinsic parameters, Disparity