Rapid Object Reconstruction for Product Augmented Reality
Overview
This project I did as part of my bachelor thesis and can be seen as my first start into the computer vision field. The download link for the thesis can be found below.
Problem Setting
Obtaining 3D models of everyday objects requires a lot of work. 3D Models are usually generated by hand using professional drawing tools, by using expensive 3D-scanners, or by using complex offline reconstruction methods that build a 3D model from a series of photographs.
Approach
Developments in 3D computer vision allow online reconstruction of specific objects with direct visual feedback and guidance to the user. This allows a user to influence and adjust the reconstruction process while the object is scanned.
In this project an application for obtaining 3D models of textured objects was implemented, inspired by the Probabilistic Feature-based On-line Rapid Model Acquisition system (ProFORMA) by Pan et al. [1]. The object is scanned with a single stationary camera by rotation around its own axis. The user, able to see the progress of the reconstructed model, can influence the further reconstruction process by appropriately rotating the object. The online modeling system consists of three main components: Tracking, Reconstruction, Visualization.
Used Technologies
This project was implemented in C++ with the use of the OpenCV library.
References
- Pan, Qi, Gerhard Reitmayr, and Tom Drummond. "ProFORMA: Probabilistic Feature-based On-line Rapid Model Acquisition." BMVC. Vol. 2. 2009.