Flying Object Detection and Classification using Deep Neural Networks

Suresh Arunachalam T, Shahana R, Vijayasri R, Kavitha T

Unmanned Aerial Vehicles, Computer Vision, Background Subtraction, Frame Differencing, Optical Flow, Edge detection, Convolutional Neural Networks.

Unmanned Arial Vehicles (UAVs) are extensively used everywhere in commercial applications such as delivering goods and medicines, taking photographs, and to monitor crowded areas. Sometimes these drones are used for capturing our private information without our knowledge. To avoid misuse of UAVs, we need to detect them in advance before entering into the protected areas. Detecting the UAV is a complex task because it is supplemented by birds, aircrafts, moving clouds, and swaying trees. To prevent this, we will detect the drones by video camera. In this paper, we compare the existing computer vision methods such as background subtraction, frame differencing, optical flow and edge detection for object detection. In our work, we will use Convolutional Neural Network for both object detection and classification to enhance its performance.

10.22161/ijaers.6.2.23