Autonomous vehicles need to navigate through complex traffic scenarios, which often involve various vehicle types, such as cars, buses, and motorcycles. Identifying the types of vehicles in proximity aids in planning appropriate traffic maneuvers, such as lane changes, merging onto highways, or yielding to emergency vehicles. The importance of vehicle type detection touches numerous aspects of modern society, including transportation, safety, and environmental concerns. This paper explores the domain of vehicle type detection within the context of deep learning and object detection. A Faster R-CNN-based deep architecture is proposed. The proposed deep model is configured to tailor the task of vehicle type detection from images. Beijing Institute of Technology (BIT) vehicle dataset is used. Experiments were conducted to evaluated proposed Faster R-CNN-based model against other models SVM, SSD, and a Simple CNN. The evaluation encompassed critical metrics, including mAP, F1-Score, ROC curves, and Average IOU. The results of experiments decisively establish the superiority of proposed model over baseline methods.