Analysis And Implication Based On Cone-mirror Catadioptric Camera Model

In the field of computer vision, Because of the small visual range, ordinary camera has limited imaging angle and gets less information. However,cone-mirror catadioptric cameras by combining with an conic reflective surface and a perspective camera.It has a larger field of view which can obtain more information of the three-dimensional space and has been widely used in robot navigation, visual surveillance and 3D reconstruction. In addition, because of low cost, easy production and simple mathematical model of projection, it has higher pixel resolution in the mirror’s periphery. Based on the analysis of cone-mirror catadioptric camera,The thesis acquire a new projection model and uses the projection relation to prove equivalence of algebraic relation between the Unitary torus model and the cone-mirror catadioptric camera model, and verifies the method by experiments. The main research contents can be summarized as follows:(1). Constrained trajectory of cone-mirror catadioptric camera’s effective viewpoint is analyzed. According to the difference of cone mirror’s angle, position of the viewpoint’s trajectory is divided into three position:i) at the bottom of the mirror; ii) on the surface of the mirror; iii) in the mirror above. Then further analyzes of the corresponding to image’s formation of cone-mirror catadioptric camera.(2). Through analyzing viewpoint and image composition of cone-mirror catadioptric camera, a new projection model is obtained by using the imaging symmetry of the cone-mirror. Through the projection relation proves equivalence of algebraic relation between the Unitary torus model and the cone-mirror catadioptric model. Then uses Unitary torus model and back projection model to obtain equation of line’s image (referred to as line-image), and through the equation to obtain line’s Plucker coordinate, line-image’s coefficients, cone’s angle, and the distance between camera with the top of the cone.(3). Achieved a number of simulated experiments and real image data using Matlab software. Verifies the feasibility and validity of the Unitary torus model and the projection relation.