The traffic surveillance and control have always been a major concern. Detecting and tracking vehicles or automobiles through satellites and surveillance cameras has made the task of tracking the defaulters and managing the vehicles on the roads a lot easier. Only detecting a moving object is not enough rather keeping track of it could possibly help in better management of traffic. So, the main objective is to detect and track the automobiles, person on the roads and provide an effective model for the same. The main challenge is to increase the detection count of the objects so that the tracking could be done efficiently and for that purpose we need to use proper detection algorithm. The dataset with around 4680 drone images has been used for the model training. After training the detection models; YOLOv5, YOLOv7, YOLOv8 the results are fed to the tracking algorithm i.e. DeepSORT. The results of the experiment reflects that the proposed object detection method has higher accuracy for faster real-time detection than the other models and the number of objects detected are also higher which helped in efficient tracking. Among all the detection algorithms YOLOv8 produced the best results with higher accuracy and maximum detections.