SBIR-STTR Award

The future of Naval surface warfare will involve a large fleet of airborne, surface, and subsurface unmanned vehicles. Unmanned surface vehicles (USV) must autonomously navigate safely through congested waters, avoiding obstacles. USVs must also be able to autonomously process their sensor data in real-time for situational awareness. UtopiaCompression’s (UC) Maritime Visual Analytics (MVA) technologies enable both military and civilian users to process visual data. In this SBIR project, UC will build on MVA to create a deep learning surface object classifier providing USVs with instant situational awareness of the objects in their vicinity as well as the unfolding events.

The future of Naval surface warfare will involve a large fleet of airborne, surface, and subsurface unmanned vehicles. As part of this plan, unmanned surface vehicles (USV) must be able to autonomously navigate safely through congested waters, avoiding obstacles. USVs must also be able to autonomously process their sensor data in real-time for situational awareness. UtopiaCompression’s (UC) Maritime Visual Analytics (MVA) technologies enable both military and civilian users to gain situational awareness through detection and tracking of surface objects. In the last decade, deep learning has become the premier algorithmic technique for classification from visual imagery, and recently UC has shown the feasibility of deep learning techniques for classification of surface objects. In this Phase II SBIR project, UC will build on its Phase I work to develop a deep learning surface object classifier (DLSOC) capable of providing Naval USVs with real-time classification of objects in the USV sensor field of view. The classifier will be integrated and tested alongside the MVA suite, resulting in a comprehensive visual analytics tool set for the Navy.