Autonomní rozpoznání dopravního značení z obrazového signálu bez řízeného osvětlení

Abstract

This paper describes the design of algorithms for autonomous detection of traffic signs without lighting control. Basis of the algorithm is marked image analysis, based on which adjustment shall be made in the image objects and their subsequent correlation with a set of patterns. Pattern analysis examines the properties of continuous objects in the image. These characteristics serve as a decision criterion for treatment sought and selects correlation patterns. The recognition system was modeled in Matlab and then implemented in C # or various Windows platforms. These platforms include Windows XP/Vista/7, as well as Windows CE and Windows Mobile. The recognition system can be run on any hardware with one of these operating systems. Speed recognition system is then given by performance and configuration of the device or computer on which it operates.