Identification of Collided Vehicles in Indian Traffic Accidents using Hierarchical Deep Learning Framework

In Intelligent Transportation Systems (ITS), accurately detecting accidents and identifying the vehicles involved is crucial for assessing accident severity. This paper focuses on the challenge of vehicle identification in accidents on Indian roads. We introduce a hierarchical detection model where an image is first analyzed for accident detection, and if an accident is detected, it is further processed to identify the involved vehicles in collision. We implemented various YOLOv8 model variants for accident and vehicle detection, finding that YOLOv8m performed best with an F1-score of 0.824 for accident detection and 0.827 for vehicle detection. The outputs of these detection units are integrated to identify collided vehicles, achieving an F1-score of 0.716 for this task.