Generating 3D Model from Video
Svetlana Mijakovska, Igor Nedelkovski, Filip Popovski · Advanced Computing An International Journal · 2014
In this paper the process of 3D modelling from video is presented.Analysed previous research related to this process, and specifically described algorithms for detecting and matching key points.We described their advantages and disadvantages, and made a critical analysis of algorithms.In this paper, the three detectors (SUSAN, Plessey and Förstner) are tested and compare.We used video taken with hand held camera of a cube and compare these detectors on it (taking into account their parameters of accuracy and repeatability).In conclusion, we practically made 3D model of the cube from video used these detectors in the first step of the process and three algorithms (RANSAC, MSAC and MLESAC) for matching data.