Information Integration and Automated Image Understanding Applied Research Programs

archived

The Office of Naval Research (ONR) is interested in receiving proposals for applied research that lead to the development of efficient, theoretically sound, and consistent algorithms for organization and fusion of high-dimensional data sources as well as automated methods for image understanding.The future naval battlespace requires an integrated sea-land-air capability in support of joint military operations potentially conducted with other services and allies. In this complex tactical picture, information gathering, processing, analysis, and understanding to enhance battlespace awareness is at the heart of the concept of joint operation. With emerging advanced technologies, information is collected in bulk from multiple data sources, either sensor-based or human-based, at a rate that exceeds the rate that humans can analyze, understand, and respond. To support highly complex and integrated naval operations with the deluge of data in a timely manner, it is imperative that certain data processing leading to situational awareness and analysis are automated to alleviate the workload for human operators and present them with only pertinent information.Within the battlespace a Naval focus is the development of the underlying computational tools to support adaptive, persistent surveillance capabilities for situational awareness, for mission planning, for force protection, etc., in the maritime domain. The maritime domain extends from 250 miles offshore to 200 miles inland, and covers a variety of environments including ocean, littoral zones, riverine environments, deserts, urban areas, etc. In adaptive persistent surveillance scenarios data is obtained from all sources, individual and networked sensors, as well as from non-sensor sources. The data may also be current or from archived databases. The network of sensor assets should be able to perform surveillance autonomously with only minimal, high-level human supervision. Therefore, the surveillance system must be able to self-coordinate and navigate autonomously, recognize and track interesting objects and agents, dwell on relevant scenes, discard irrelevant data, index and store data of interest for forensic investigations and after action analysis, and provide timely alerts with pertinent information to warfighters.

High