Robust lane detection in hilly shadow roads using hybrid color feature

Over the past two decades, autonomous vehicles for consumers have emerged as an imperative area of research. For driver assistance systems, the lane detection technique forms a key element to augment safe driving and afford warning in case of hazard. In this article, the difficulty of detecting lane amidst shadows and strong lighting condition is addressed which is significant for driving automation on mountainous roads. The proposed approach is tailored to mitigate the shadow effects cast on road scenes by using a hybrid saturation feature which in turn would meet the need for robust lane identification. Computer vision techniques are used to extract the existence of visual cues and label the lanes in the camera frames. Experimental analysis is carried out on real road hilly images using the proposed approach.