Transportation serves as a vital link for our economy, various types of vehicles have been an integral part of economic growth but have led to an enormous increase in traffic and its attendant problems like accidents and mortality. Hence, there is ample scope for improvement of newer technologies that have aimed to reduce the negative impact of high volumes of high-velocity traffic, easing traffic congestion and making transport safer. The objective of this paper is to review some of the recent advances in the field of lane detection in autonomous vehicles, the problems perceived, and the solutions designed. Image processing technologies used in traditional lane detection systems were observed to have inherent disadvantages. Further advances in lane detection consist of models built to detect lanes in complex environments mainly using two-stage lane feature extraction implemented with the you only look once (YOLO) v3 algorithm for precise and faster robust lane detection in complex real-life scenarios. In the next few years, huge breakthroughs will happen in autonomous vehicles being able to systematically detect lanes without supervision, using new computer vision-based lane detection methods to make complete autonomous vehicular travel a safe reality.
Recent Advances in Computer Vision Technologies for Lane Detection in Autonomous Vehicles
Algorithms for Intelligent Systems
International Conference on Computer Vision and Robotics ; 2023 ; Lucknow, India February 24, 2023 - February 25, 2023
28.09.2023
16 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
T-S YOLO v3 You Only Look Once (YOLO) , Random sample consensus algorithm (RANSAC) , Lane localization , Convolutional neural networks (CNN) , Deep Q learning (DQLL) , 3D network models , Adaptive learning methods , Lane detection , Temporal , Lane boundary regression , Semantic segmentation network , Recurrent neural network (RNN) , Region-based convolutional neural network (RCNN) , Spatial , Deep reinforcement learning Signal, Image and Speech Processing , Cyber-physical systems, IoT , Control, Robotics, Mechatronics , Engineering , Computer Imaging, Vision, Pattern Recognition and Graphics