A Study of Detection and Tracking of Artificial Intelligence in UAVS Using Machine-Learning Approach

Abstract

Detecting multi-rotor unmanned aerial vehicles is one of the article's goals, and it will be tested in this article (UAVs). Many sectors, such as the protection of fragile structures or the preservation of privacy, need such a job. Computer vision techniques such as the Oriented FAST and Rotated BRIEF (ORB) feature detector were used to build our system initially. As an alternate detection method, the machine-learning approach was employed due to the poor success rate in real-world circumstances. Using the "Common Objects in Context dataset," 1000 samples of UAVs from the Safe Shore dataset were added to the preset dataset. Drones in a static picture and films showing a drone in the sky have been shown to be successful and reliable by four basic experiments—one with a separate flying item in the sky, and three others using several drones together in one image. In ideal circumstances, the detection success rate was 97.3%.