Video Object Detection Segmentation Using Yolov8 Algorithm for Autonomous Driving

Abstract

Smart and autonomous vehicles are systems that reduce or do not require human intervention in the face of unexpected and changing situations in life. In order for autonomous vehicles to intervene in the mentioned variable situations, the variable must first be detected. Various object detection studies are used to detect these variables. Nowadays, studies on autonomous vehicles and object detection are being carried out and they continue to be developed every day. In this study, video object detection segmentation for autonomous driving was performed using the Yolov8 algorithm. In the study, the CamVid (Cambridge-Driving Labeled Video Database) data set found in the literature was used. The study consists of a total of 700 images and 21 classes. As a result of the study, it is seen that the classes "person", "car", "traffic light", "potted plant" and "bicycle" have the best detection results. © The Institution of Engineering & Technology 2024.