Real-time line matching based speed bump detection algorithm

Abstract

Active safety is one of the main technologies in the modern automotive industry. Active safety is achieved with advanced sensors that employ Computer Vision algorithms. To assist the driver and eventually to achieve fully autonomous driving, these sensors must be able to perform functions like lane detection, traffic signs recognition, detection of dynamic obstacles and irregularities of the road such as speed bumps. Speed bump detection is an important function for an advanced driver assistance system. In this paper, we present a speed bump detection algorithm that uses a mono-camera sensor as an input. We propose a real-time robust speed bump detection algorithm that first identifies a region of interest, based on semantic segmentation, and later on applies a novel variant of ED Line algorithm.