Drone Detection Based on FD-HOG Descriptor

Abstract

The rapid development of Unmanned Aerial Vehicle (UAV) technology, also known as drones, has made people benefit in many ways, but it also brings privacy and security issues, such as appeared at private place, airports, prisons, etc. Therefore, the detection of drones in a specific area is crucial. Video detection is an effective method with various advantages. In this paper, we have used the background subtraction method to detect the moving object in the video sequence which recorded by static cameras, then extracted the global Fourier descriptors and the local HOG features of the moving object images. Finally, the FD+HOG features have been sent to the SVM classifier for classification and recognition. Our algorithm is simple and efficient, an overall with 98% accuracy was obtained in our data set.