Malicious UAV detection using integrated audio and visual feature

Date : Aug 2021 - Jul 2022

<div><p>Mini drones which are also known as unmanned aerial vehicles (UAVs) become popular now a day. It is used in different sectors such as Agriculture, road monitoring in foggy weather, and also different industries for different applications. As their use increase, they pose serious security threats to public privacy and other sectors including security agencies. By keeping in mind the current need for security services we chose our project to assist the presently required field.The project is based upon Digital image and Audio processing using machine learning and deep Neural networks. The detection of a drone is an open research issue for security companies. Several schemes have addressed this issue. However, each plan has some limitations. Such as High demands under extreme weather conditions and dataset. This project proposes a novel framework consisting of hybrid handcrafted, and deep neural networking to detect and identify malicious drones based on audio and visual information. The dataset contains the sound of different birds, planes, thunderstorms, and images, with different resolutions. Various kernels of Support Vector Machines (SVMs) are used to classify features. Currently, it gives an accuracy of about 98.3%.</p></div>