The essence of object detection in Computer Vision is to identify and precisely locate an object in a scene. In transportation management, it detects traffic signs for application in autonomous vehicles and other road users. Traffic objects and the scene must also be documented and monitored for efficient traffic management. The signs are prone to damage and occlusion, which may compromise road safety, thus the need to detect, document, and monitor traffic signs. This paper proposed a traffic signs detection and recognition model using the state-of-the-art YOLOv8 model and an inventory management system for keeping traffic records. We also proposed a monitoring technique by viewing traffic scenes uploaded on customized Google My Map and Google Map, made possible by geolocation information of traffic signs. Our trained model achieved 87.1% mAP50 and 61.2% mAP50-95 on our dataset. We compared the performance of our model to other versions of YOLOv8.
Viewing on Google Maps Using Yolov8 for Damaged Traffic Signs Detection
Communic.Comp.Inf.Science
International Conference on Technologies and Applications of Artificial Intelligence ; 2023 ; Yunlin, Taiwan December 01, 2023 - December 02, 2023
28.03.2024
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch