Published Paper Details:

Vehicle detection through satellite video imagery has witnessed remarkable advancements, revolutionizing remote sensing and Earth observation. In this survey research, we explore the latest methodologies, approaches, and innovations in satellite video-based moving vehicle detection, with a special focus on the Semantic-Embedded Density Adaptive Network (SEDAN) architecture. Satellite videos provide an unparalleled perspective of Earth's surface, enabling us to grasp dynamic events and human activities. However, discerning moving vehicles within these videos presents significant challenges, including low image quality, varying lighting conditions, and diverse backgrounds. Traditional handcrafted features and shallow learning algorithms often fall short in delivering robust results under such circumstances. In response, deep learning-based solutions have emerged, with the SEDAN framework emerging as a promising approach. Drawing insights from a meticulous review of 22 papers, this survey research delivers a comprehensive overview of vehicle detection using satellite video data, highlighting the transformative potential of satellite technology in remote sensing and its crucial applications in various domains, including disaster management, traffic monitoring, and national security.