Phone: (+61 8) 6488 4775
Fax: (+61 8) 6488 1089
Real Time Object Recognition Using RGB-D Video Data
3D object recognition has gained significant attention from the research community over the past few decades. The goal of object recognition is to determine the identity of an object being observed in a scene from a set of known objects in a database. Automatic real-time object recognition will be a key feature of future generation robots. The focus application of these robots is to provide assistance to the elderly and people with disabilities. Other applications include underwater exploration, rescue operations, maintenance operations for oil and gas applications, or military operations such as cave clearing, search for underground tunnels or remote monitoring. Despite decades of research in robot vision, the performance of existing state of the art recognition systems does not allow a mainstream deployment of autonomous robotic object recognition systems in real-world environments. Occlusion, clutter and variations in illumination limit the recognition performance. Moreover, existing object recognition systems operate primarily in controlled environments, e.g. plain background scenes. This research will explore automatic 3D object recognition in real time under clutter and occlusion. The proposed study will use improved image features and techniques to accurately recognize 3D objects in a scene in real time.
1. Assisted living (for elderly & people with disabilities)
2. Mining Industry
3. Manufacturing Setup
4. Security & Surveillance