Target Identification and Events Analysis of UAV Based Image using Deep Learning

Abstract

Deep learning is used to learn, track, and discover targets from the data obtained. Deep learning is also used in target identification systems to improve the ability of these systems to identify the position of the targets. Unmanned aerial vehicles (UAVs) ?�? also known as drones ?�? with integrated deep learning can patrol border areas, identify potential threats, and transmit information by analyzing events based on deep learning about these threats to response teams or trigger alarms. Generally, the conditions listed above were what motivated me to deep dive into this research study on event analysis and target identification of UAVs based on deep learning. To contribute valuable research by integrating deep learning and machines like UAVs to solve the problem that happened with the limitation of Intelligence, Surveillance, and Reconnaissance (ISR) operations by human beings, especially in remote and complex areas. Activity that happens in every moment is tedious and complex for a human to analyze, but in this paper, we presented target identification and event analysis of UAVs based on deep learning methods to analyze events continuously from an aerial view using deep learning. The main purpose of this study was to review related works, train publically available aerial datasets and self-collected datasets from UAVs using pre-trained models, and present the result of the best models that performs event analysis, target identification, and classification in real-time, with computation power requirements that can be met with a common laptop computer and android phone contain above android version 11.