Fredrik Kahl is a Professor in image analysis and computer vision, and he finds that the use of robot computer vision today is limited to very specific scenarios or controlled laboratory environments.
“Examining today’s best-performing systems for visual processing reveals that problems such as object recognition and 3D scene reconstruction have largely been studied independently and there is no single, integrated modelling framework. For instance, object recognition and 3D scene reconstruction, despite being strongly connected problems, are treated independently and an integrated theory is lacking”, says Fredrik Kahl.
This is what the project Semantic Mapping and Visual Navigation for Smart Robots is about, which has recently received funding from the Swedish Foundation for Strategic Research, under the program Smart systems.
“In this interdisciplinary project, we rely on expertise from computer vision, machine learning, automatic control and optimization in order to take the current state-of-the-art in autonomous systems to the next level of perception, cognition and navigation, and towards key capabilities of robots able to effectively act in the real world”, says Fredrik Kahl.
Computer vision is about developing automated systems to analyze and understand digital pictures, from for example cameras. Machine learning involves developing algorithms that gives the system the ability to make decisions based on historical data and real-time data. In robotics research, systems are developed to use feedback and learning to interact with their environment and perform tasks autonomously or in cooperation with people.
Despite the fact that in recent years there has been a dramatic development within all three areas, today's systems only solve partial problems, with one specific method.
“We believe that in order to reach further, it is necessary to develop smart systems that are capable of integrating the different aspects of vision in a collaborative manner”, concludes Fredrik Kahl.
For demonstration, the project will develop an autonomous system for the visual inventory inspection of a supermarket using small-scale, low-cost quadcopters. The system will provide a complete solution for visual navigation and 3D mapping where not only scene geometry is modelled, but also semantic constraints are integrated. The research is relevant for many industrial applications, such as self-driving cars, unmanned ground vehicles, scene modelling and inspection in general.
The project has gathered expertise from computer vision, machine learning, automatic control and optimization. The project is led by Professor Fredrik Kahl from the Department of Signals and Systems, at Chalmers University of Technology. He has been one of the pioneers in developing modern optimization techniques for applications in computer vision. For this, he was awarded the ICCV 2005 Marr Prize which is considered the most prestigious prize in computer vision. He is the only Scandinavian to have received the award. He has also received support from the European Research Council (ERC starting grant) and the Swedish Foundation of Strategic Research (Future Research Leaders).
The management of the project also includes Professor Cristian Sminchisescu and Professor Kalle Åström, at the Centre for Mathematical Sciences, Lund University, and Anders Robertson, Professor of Automatic Control at Lund University.