The human eye, connected with the human mind, is the base tool that allows us to perceive and interpret the world surrounding us. Our eyes let us not only simply see objects- we can recognize forms, follow movements, estimate distanoes and velocities. The computer vision is the discipline, which tries to apply those capabilities to computers. This allows to improve and complete our vision and, on the other hand, to eliminate the need of human interaction in certain applications. The number and importance of applications of computer vision in to- day's world is difficult to underestimate. Surveillance, robotics, assistance in surgical operations and CAD / CAM systems are only some in the wide range of disciplines which are developed mostly thanks to progress in computer vision_ The reasons for applying computer vision are numerous: its flexibility, ability of fast treatment of large amounts of data and, above all, the possibility to exprem information in a clear and intuitive way. One branch of computer vision deals with the extraction of threedimensional information from images. The ability of interpretation of 3D data is necessary in some kinds of applications. Consider for example robotics, the problem of obstacle avoidance, or applications for 3D model visualisation. In this work we are interested in using computer vision to construct threedimensional models of scenes. Reconstruction of geometric models allows us to observe scenes from different points of view. Texture information, which can be extracted from images, permits to obtain photorealistical renderings of models.