Rapid Reconstruction Technique Of Visual 3D Model Using Scaled Window Phase-Only Correlation (SWP-OC) On The Sequence Of Images Captured Surrounding The Object

Mohd Saad, Wira Hidayat and Mohd Saad, Norhashimah and Abd Latif, Mohd Juzaila and Abd Manap, Nurulfajar and Hasan Basari, Abd Samad and Ahmad Radzi, Syafeeza (2017) Rapid Reconstruction Technique Of Visual 3D Model Using Scaled Window Phase-Only Correlation (SWP-OC) On The Sequence Of Images Captured Surrounding The Object. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text Rapid Reconstruction Technique Of Visual 3D Model Using Scaled Window Phase-Only Correlation (SWP-OC) On The Sequence Of Images Captured Surrounding The Object.pdf - Submitted Version Restricted to Registered users only Download (510kB)

Abstract

This project demonstrates the use of vision camera to capture a sequence of 2D images surrounding the object to reconstruct the 3D model of the external view of the object.3D reconstruction technique with a realistic rendering process can normally be applied either by using a sequence of images captured by using a typical camera or by using a stereo camera.By using a sequence of images surrounding the object give an advantage over stereo image because it allows the object to be modelled completely.One of the best techniques of 3D reconstruction algorithm that has been verified with stereo image is Scaled Window Phase Only Correlation (SW-POC).This technique exhibits accurate and robust correspondence matching in both shortand wide-baseline stereo pairs with a very low computational cost.The proposed method employs the POC-based correspondences matching with the concept of Scaled Window Matching (SWM) to find the accurate correspondence from a short and wide baseline stereo image pair.POC technique used phase information obtained from Discrete Fourier Transform (DFT) of images.Due to that reason,the disparity can be limited to horizontal direction only to rectify the stereo image pair.This is one of the key points that help in reducing the computational time for this method.

Actions (login required)

View Item