Özet:

Abstract In the modern era, accurate vehicle detection and classification has become one of the key challenges due to the rapid increment in the number of varied size vehicles over the roads, particularly in urban regions. There have been explored and implemented promiscuous models for vehicle detection and categorization in the last few years. However, existing advanced frameworks have several issues related to the correct identification of the vehicles due to the shadow problem, which results in lower identification accuracy for prioritization of emergency vehicles. Further, these models consume more time in execution as well as intricate to implement and maintain in real-time. In this research work, a new framework for vehicle detection and classification has been proposed. This novel framework is based on Yolo-v7 and a Gradient boosting machine (GBM) to prioritize emergency vehicles in a faster and more accurate manner. For alleviating huge traffic as well as safety concerns, this suggested framework centers on the accurate identification of vehicles class in intelligent transportation system (ITS) for assigning priority to emergency vehicles to get a clear path. The findings of the suggested framework are highly optimal as well as enhanced in comparison to the previous models. The evaluated performance metrics i.e. accuracy, precision, recall, and F1-score of the suggested enhanced Yolo-v7 and GBM-based model are 98.83%, 96%, 97% and 98%respectively. This proposed research work can be extended in the future for more accurate vehicle identification to handle multifarious challenging circumstances, namely snow and rainy conditions or during the night.

Anahtar Kelimeler: