Codesign methodology for computer vision applications

Abstract

The work presented in this paper has been performed under a Spanish research project. The main aim of the tasks, we were responsible of, was the development of a vision subsystem for 2D image preprocessing. These algorithms are the first step of a 3D reconstruction algorithm. These algorithms have been improved by the addition of fault tolerance capabilities. To achieve this goal, the classical codesign methodology has also been improved to allow for the fault tolerant system codesign. This paper presents the results obtained from this work.