engleski

MODEL FOR APPLICATION OF OPTICAL PASSIVE SFM METHOD IN RECONSTRUCTION OF 3D SPACE AND OBJECTS

Reconstruction of objects and space-based on images is the topic of active research in a recent decade. The reason for that is because high-quality reconstruction is hard to achieve, but if achieved successfully there is a wide range of possible applications. Particularly in disruptive technologies such as virtual reality and augmented reality. Advancement in computer and optical hardware has enabled development of acceptable reconstruction that can be applied for some type of purposes. Depending on the goal of reconstruction there is a wide variety of possible approaches and methods. This thesis is based on structure from a motion approach that can be used for reconstruction of a single object, room, building, street, or city. The method uses a set of images that have a targeted object or space for achieving reconstruction. Structure for motion is based on feature extraction, camera registration and stereo vision. The method is not designed for real-time reconstruction.

Structure from Motion ; SFM ; 3D reconstruction ; Feature extraction ; SIFT ; KAZE ; Stereo vision

nije evidentirano