Recognition of arrows in the environment of road markings

In the domain of image understanding in road environment, we have developed a method to recognize and locate arrows. The ambiguities in natural environment and the execution time for online processing motivated our approach. The model of the scene (including road, vehicle, camera and markings) is analysed and applied as early as possible in the appropriate levels of processing. We also tried to simplify the algorithms which work on huge data in the early stage of processing. In this article we present the model of the scene and its intervention in the processing as well as the algorithms.