Future driver assistance systems will rely on a variety of sensors, including camera-based systems, to achieve an optimal perception of a vehicle's surroundings. With increasing resolutions and the inevitable real-time requirements of these safety-relevant systems, multi-camera setups, demand processing performances and data rates that cannot be satisfied by current standard PC components. To meet these requirements, this paper presents a flexible FPGA-based design approach for a central onboard image preprocessing unit, which offloads computation intensive tasks to a dedicated FPGA accelerator platform with support for real-time execution. This paper is organized as follows. The section 2 gives an overview of related research. In section 3, a brief overview of the basic architecture is given, followed by an analysis of its usability for image processing in driver assistance functions. Section 4 describes the integration of the architecture in a concrete driver assistance scenario. The motion estimation algorithm and its implementation in hardware is presented in section 5 followed by benchmark results in section 6. A final conclusion is given in the last section.
Real-time image processing for camera-based driver assistance applications
Echtzeit-Bilderkennung zum Einsatz in Fahrerassistenzsystemen auf der Grundlage von Kamerasystemen
2010
10 Seiten, 5 Bilder, 12 Quellen
Conference paper
English
Real time image processing for camera based driver assistance applications
Automotive engineering | 2010
|Real-Time Camera Link for Driver Assistance Applications
British Library Conference Proceedings | 2009
|Real-time camera link for driver assistance applications
Tema Archive | 2009
|