Hierarchical Scheduling for Oil Spill SAR Image Detection in Space-Air-Ground Integrated Networks

Oil spills represent significant environmental hazards in ocean ecosystems, requiring rapid and accurate detection and response mechanisms. Due to its efficacy, synthetic aperture radar (SAR) is an important tool for remote monitoring of ocean environments. However, operational constraints such as communication time between low Earth orbit (LEO) satellites and ground stations (GS), as well as finite buffers, pose challenges. To address these issues, this paper proposes a space-air-ground-integrated network (SAGIN) scenario comprising LEO satellites, high altitude platforms (HAPs), and ground stations (GSs), suitable for rapid and accurate oil spill detection using SAR images. In addition, a novel three-step algorithm is proposed to facilitate SAR image transmission effectively. The first step of the proposed algorithm focuses on maximizing the transmission of ocean SAR images from the LEO satellites to the HAPs. This optimization involves the neural Myerson auction (NMA)-based scheduler, a second-price auction mechanism with deep learning methodologies for truthfulness, distributed computation, and low computational complexity. In addition, the pathloss is also considered for effective communication system design. The second step algorithm aims to maximize data transmission efficiency from HAPs to GSs, employing scheduling strategies based on the Markov decision process (MDP). This algorithm enables optimal decision-making over time steps, considering the throughput of each GS and the age of the information (AoI) ratio. Finally, an image compression rate control algorithm based on Lyapunov optimization is employed at the GSs to enhance detection accuracy while ensuring queue stability. The performance of the proposed three-step algorithm is intensively evaluated against various benchmark algorithms, demonstrating its superiority in oil spill detection efficacy.