3D Reconstruction Using Silhouette Coherance For Image Acquisition Under Circular Motion

Norliana, Azhar (2015) 3D Reconstruction Using Silhouette Coherance For Image Acquisition Under Circular Motion. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) 3D Reconstruction Using Silhouette Coherance For Image Acquisition Under Circular Motion.pdf - Submitted Version Download (412kB)

Abstract

This thesis presents the reconstruction for 3D image by using silhouette algorithm and camera calibration in order to enhance the 3D reconstruction. This technique is useful to reconstruct an object’s geometry taken under circular motion or turn table. Once the image acquisition steps is finish, the camera calibration is done in order to recover the extrinsic and intrinsic parameters of camera. The silhouette-based approach is used in this thesis. Silhouette coherence is a generalization of epipolar tangency constraint used to calculate the motion from silhouette or the outline of the object. This approach is capable of exploiting the information in the silhouette compared to the earlier silhouette-based approach that only depends on the epipolar tangency. With the flexibility of this technique makes it usable in a situation where outer epipolar tangents or point correspondent are not available. This thesis also presents the algorithm that used for exploiting silhouette coherence for camera motion estimation. The proposed algorithm is also used to reconstruct a high quality 3D model under circular motion sequence even when the silhouettes are truncated or the epipolar tangency points are not available. The sequence of images used in this thesis were obtained by using turn table. Visual evaluation is done for the final result of the 3D image for different camera angle and different number of sequence. All the factors that cause effect to the final 3D image are evaluated in order to achieve better 3D reconstruction result.

Actions (login required)

View Item