A systematic review: computer vision algorithms in drone surveillance

<p class="p1">Drones have emerged as advanced Cyber-Physical Systems (CPSs) with significant potential in data collection and environmental monitoring. Their ability to operate via wireless communication channels makes them integral to various IoT applications, such as surveillance, delivery services, traffic monitoring, and precision agriculture. Drones surpass traditional surveillance methods by offering better mobility and broader coverage, enabling efficient decision-making in diverse contexts, However, object detection in drone-captured images poses challenges due to varying spatial resolutions, the large number of objects, and their diverse sizes in aerial imagery. This paper provides a comprehensive review of drone-based surveillance techniques, focusing on object detection and tracking algorithms, relevant datasets, and exploration strategies. By analyzing current methods and identifying key trends, this study aims to highlight advancements and opportunities for improving the performance and reliability of drone-based surveillance systems. <p class="p1"> <p class="p1"><strong>Received on, 23 December 2024 </strong> <p class="p1"><strong>Accepted on, 19 February 2025 </strong> <p class="p1"><strong>Published on, 03 March 2025</strong>