Autonomous Vehicle Object Detection

Abstract

Autonomous vehicles require a reliable perception system to understand objects present in their surrounding environment. Object detection is therefore an important component of autonomous driving and Advanced Driver Assistance Systems (ADAS). This paper presents a software-based real-time road-object detection system using Python, OpenCV, and YOLOv8. A webcam provides live video input, OpenCV performs frame capture and basic preprocessing, and YOLOv8 detects and classifies objects such as cars, buses, trucks, motorcycles, bicycles, and pedestrians. The system displays bounding boxes, class labels, and confidence scores for the detected objects. The proposed work focuses on the visual-perception stage of autonomous vehicles and is designed as a practical prototype that can run on a conventional laptop or desktop computer. The methodology covers problem identification, literature survey, system analysis, system design, implementation, testing, and result analysis. Standard metrics including precision, recall, F1-score, mean Average Precision (mAP), and frames per second (FPS) are defined for evaluation. The paper also discusses the limitations of camera-only detection and identifies future extensions such as lane detection, object tracking, distance estimation, traffic sign recognition, collision warning, adverse-weather handling, and edge deployment.