PROJECT TITLE :

Deep Learning in Lane Marking Detection A Survey

ABSTRACT:

The detection of lane markings is a step that is fundamental to intelligent driving systems but also extremely important. Not only can it assist with vehicle positioning and forehead car detection, but it can also provide pertinent information about the road conditions to prevent drivers from leaving their lanes. The detection of lane markings, on the other hand, is plagued by a number of obstacles, such as extreme lighting, absent lane markings, and obstructions from obstacles. In recent times, deep learning-based algorithms have been attracting a lot of attention in the intelligent driving society due to the excellent performance they offer. In this paper, we review deep learning methods for lane marking detection. Our primary focus is on these methods' network structures and optimization goals, which are the two most important factors that determine the level of success achieved by these methods. In addition, we provide a summary of the existing datasets, evaluation criteria, and common data processing techniques that are related to lanes. We also compare the detection performance and running time of a variety of methods, and we conclude by discussing some current challenges and potential future trends for an algorithm that detects lane markings based on deep learning.