Remote Video Surveillance using LEO Satellites Communications

Video surveillance systems have rapidly expanded, driven by their critical roles in security and traffic monitoring. This expansion has produced vast data volumes, causing bottlenecks in communication systems due to the time-sensitive and bandwidth-intensive nature of surveillance data. To address these challenges, research has increasingly focused on developing algorithms to compress redundant data effectively, evolving from traditional methods to advanced neural video compression techniques. These methods aim to transmit minimal data without sacrificing quality. Our approach utilizes the semantic communication paradigm, emphasizing the transmission of data’s meaning rather than the data itself, to enhance efficiency in low-bandwidth environments. We propose a semantic compression framework specifically for video surveillance that employs novel object detection models and varying levels of semantic abstraction. This framework is designed to convert surveillance footage into compact, meaningful representations, optimizing data transmission in scenarios like remote surveillance via satellite.

Faculty Supervisor:

Lokman Sboui

Student:

Partner:

École Supérieure Privée d'Ingénierie et de Technologies (ESPRIT)

Discipline:

Engineering

Sector:

Information and Communications Technology; Artificial Intelligence; Environmental Science and Technology

University:

École de technologie supérieure

Program: