Vehicle Tracking System in Drone Imagery with YOLOv5 and Histogram

In this study, we propose a vehicle tracking system targeting drone footage. The proposed system utilizes the real-time object detection network, YOLOv5, to acquire vehicle location information and segment the vehicle regions based on it. The system analyzes the histogram of the segmented regions, compares them with past frames, and determines whether the objects are identical to perform tracking. To enhance the efficiency of histogram comparison, the algorithm is designed to compare objects only within a certain radius using coordinate information and past frame object data. The MOTA (Multi Object Tracking Accuracy), a representative tracking evaluation metric, showed 90%. However, it is important to consider the limited environment of data usage and experiments. The results of this study suggest that the real-time performance of the vehicle tracking system can be utilized in various fields such as traffic control, vehicle management, and accident response.