Super-resolution modeling

Abstract

Recently, many virtual reality and robotics applications have been called on to create an environmental model from real scenes, because virtual views can be freely created from the model. However, the resolution of sensor limits that of created virtual views. In this paper, we propose super-resolution modeling concept that estimates a higher resolution texture on object from an image sequence. This concept is suitable for an environmental modeling for virtual reality applications. We also propose a modeling system with omnidirectional image sensor, and estimate the system to confirm the effectiveness of super-resolution modeling.