UW-SVC: Scalable Video Coding Transmission for In-network Underwater Imagery Analysis

Abstract: Underwater imagery has enabled numerous civilian applications in various domains, ranging from academia to industry, and from industrial surveillance and maintenance to environmental protection and behavior of marine creatures studies. The accumulation of litter and plastic debris at the seafloor and the bottom of rivers are extremely harmful for the aquatic life. We propose a solution for this problem using a team of Autonomous Underwater Vehicles (AUVs) to exchange the recorded video in order to reconstruct the seafloor regions of interest. However, underwater video transmission is a challenge in the harsh environment in which radio-frequency waves are absorbed for distances above a few tens of meters, optical waves require narrow laser beams and suffer from scattering and ocean wave motion, and acoustic waves, while long range, provide a very low bandwidth and unreliable channel for communication. In our solution, the scalable coded video of each vehicle is shared with a selected group of receiving videos, pseudo-multicasting, through the acoustic channel. Presented evaluations, including both simulations and experiments, confirm the efficiency and flexibility of the proposed solution using acoustic software-defined modems.