LiDAR-Based Advanced Driver Assistance for Vehicles

Safety has been a primary concern in automotive systems. The advanced driver assistance systems (ADAS) aid in reducing accidents caused by driver error, improving traffic safety, and enhancing driving efficiency. The use of sensors is the most popular option for autonomous vehicles among the many that have been proposed in recent years. Some of the important sensors used in ADAS include Camera, Radar, LiDAR and ultrasonic sensors. A Light Detection and Ranging (LiDAR) has one of the most popular stereo-vision technology. The region of interest (ROI) on images is found using object segmentation through LiDAR cloud points. The region of interest image will be fed into YOLOv4 and YOLOv7 for evaluation. In this paper, LiDAR is used to detect the target vehicle using YOLOv4 followed by YOLOv7 to improve the performance of detection. The PandaSet is the LiDAR dataset used in the paper.