ASSESSMENT OF VISIBILITY IN COMPLEX ROAD SCENES USING DIGITAL IMAGING

摘要

In order to assess the impact of a lighting system on the visibility of objects in a complex road scene, a classic approach consists in selecting objects of interest to compute their luminance, the iuminance of their background, and then their visibility level. Implementing this approach to analyze calibrated digital images is not easy, mostly because of the necessity to arbitrarily delimit backgrounds and to compute the luminance of non uniform regions. In this paper, we present a method to assess of visibility in digital images of complex road scenes without prior knowledge or assumption concerning the objects in the scene and their luminance, or that of their background. The idea is to use a segmentation technique to detect the edges which make the objects visible, and to compute the local contrast along these edges. The results give a clear picture of the impact of one lighting condition compared to another. In this paper, we present the image processing technique, along with sample results.