Abstract
This study proposes an integrated framework for efficient traffic object detection and classification by leveraging advanced deep-learning techniques. The framework begins with the input of video surveillance, followed by an image-acquisition process to extract the relevant frames. Subsequently, a Faster R-CNN (ResNet-152) architecture was employed for precise object detection within the extracted frames. The detected objects are then classified using deep reinforcement learning, specifically trained to identify distinct traffic entities, such as buses, cars, trams, trolleybuses, and vans. The UA-DETRAC dataset served as the primary data source for training and evaluation, ensuring the model’s adaptability to real-world traffic scenarios. Finally, the performance of the framework was assessed using key metrics, including precision, recall, and F1 score, providing insights into its effectiveness in accurately detecting and classifying traffic objects. This integrated approach offers a promising solution to enhance traffic surveillance systems and facilitate improved traffic management and safety measures in urban environments.