Enhancing Vehicle Feasibility Testing with YOLO: A Computer Vision Approach

Vehicle crashes, resulting in over 1.19 million fatalities in 2023, highlight the critical need for better road safety measures. Current vehicle inspection systems, which rely heavily on manual assessments, are time-consuming, error-prone, and inefficient. This research develops an automated vehicle feasibility assessment system using the YOLOv11 segmentation algorithm to detect key vehicle components, including headlights, brake lights, turn signals, hoods, windshields, and wheels. By leveraging standard CCTV infrastructure, the system provides a cost-effective, real-time alternative to traditional methods. Three YOLOv11 variants, yolo11n-seg, yolo11s-seg, yolo11m-seg were evaluated, with results showing yolo11m-seg achieving the highest precision, yolo11s-seg excelling in recall, and yolo11s-seg being suitable for real-time use in resource-constrained environments, offering a effective solution to improve vehicle inspections and road safety.