An Effective and Secure Transfer Learning-based Mechanism for Drone Detection

The rapid growth of drones and UAVs have brought significant benefits to various industries such as surveillance, agriculture, logistics, and defense. However, this growth has also led to increased security and privacy concerns, thus some capable techniques are required which can detect drones in restricted areas. To meet this goal, an effective approach to drone detection using transfer learning techniques is presented in the paper. The proposed mechanism also provides a secure mechanism for data transfer between drones and the ground station servers. The details of the proposed scheme include dataset utilized and methodologies used, followed by a detailed description of the utilized transfer learning models, like, DenseNet121, VGG19 and MobileNetV2. Last, the results are shown based on standard performance parameters i.e. accuracy, recall, precision and F1-score that have been obtained from the implementation. During the comparison, it has been observed that the proposed mechanism outperformed the other existing techniques.