Efficient vehicle detection and accurate vehicle tracking are gaining significant improvement in the field of traffic management. However, due to less lighting, occlusion, and complex backgrounds, their precise detection and tracking are challenging. To address these issues, this paper proposes a vehicle detection and tracking framework called DeepYOLO, which uses a computer vision-based vehicle detection system using You Only Look Once (yolo-v8) along with a Deep Learning-based Simple Real-Time Tracker (DeepSORT). The video footage of vehicles captured from the cameras installed at the traffic intersections will be initially converted into multiple frames and then given to the yolo-v8 detector. The detected frames will be then inputted to the DeepSORT tracker to effectively track, count, and to find the distance traveled by each vehicle at the junction using the bounding box coordinates. The proposed model is trained with MS COCO 2017 dataset and shown good results when compared with the different versions of yolo and faster R-CNN detection models.