Multi-frame Fusion Methods Based on Cepstral Coefficients for Drone Classification

Authors : Nida Kumbasar, Rabiye Kılıç, Emin Argun Oral, Yücel Özbek

Pages : 892-916

View : 94 | Download : 205

Publication Date : 2025-12-31

Article Type : Research Paper

Abstract :The increasing popularity of drones in recent years has resulted in privacy and security vulnerabilities. Today, drones can be easily purchased and used. Therefore, people can take advantage of these drones to intrude into private areas. Detecting and identifying the presence of drones in an area is of great importance. There are different detection techniques, such as video, sounds, thermal imaging, and Radio Frequency (RF) signals, in drone detection and classification. In this study, RF signals are used to classify a drone. In order to effectively classify drones with high performance, the multi-frame majority voting method is recommended using the cepstral coefficients. For this purpose, drone signals are divided into multiple frames (2, 4, and 8), and each frame is extracted with Mel Frequency Cepstral Coefficients (MFCC) and Linear Frequency Cepstral Coefficients (LFCC) attributes. Then, each frame is classified by Support Vector Machine (SVM), and the predictions obtained from the frames of a drone signal are subjected to majority voting. Results were obtained with 100% accuracy for drone classification (4-Class) and 99.11% accuracy for defining operating mode (10-Class). The proposed method outperforms existing methods in drone classification using the DroneRF dataset.
Keywords : Drone classification, RF signal, MFCC, LFCC, SVM

ORIGINAL ARTICLE URL