Geometric models for video surveillance in road environments: vehicle tailgating detection

Abstract

Traffic accidents constitute one of the main causes of death in many countries. Despite the current efforts devoted to mitigate the effects of road incidents, there are still some variables affecting this problem which are not yet under control or regulation. Spain, for instance, still lacks official regulations about especially risky driving behaviours, such as tailgating. In many cases, the rationale behind is that these behaviours are hard or expensive to detect reliably, thus limiting the extent of the automatic detection systems. This paper proposes a method to identify certain elements in road scenarios, define geometric models that allow computing quantitative measures of the scene and, consequently, detect offending driving behaviours. In this work, we have focused on the particular case of study of tailgating detection. However, the proposed geometric models might become the basis of many other useful applications.