Asymmetric Threat Assessment Using Electro-Optical Image Systems

Asymmetric threat assessment from military platforms, including early detection and classification by electro-optical means, is a complicated matter. These threats can be for instance explosives-packed rubber boats, minecarrying swimmers and divers in a marine environment or terrorists, improvised explosive devices and small vehicles in land or urban environments. Detection is hampered by the threat optical characteristics, as well as the local environment, such as atmos here and background, which impact on contrast and signal to-noise ratio. This is true for applications on board of naval platforms in coastal environments, at open sea, for the present-day threat of piracy and observation of hostile activities from the air as well as from land. Background variability as well as target variability are important issues to address. These lead to large amounts of spatial and temporal clutter, which need to be taken into account in the detection and classification processing. For detection as well as classification in visible light and infrared imagery, it is essential to have efficient processes that make use of sophisticated clutter rejection schemes. To have large spatial coverage, because threats can appear from everywhere, a wide field-of-view search system is required. The detection process is performed on the search system, in which clutter rejection is adapted to the environmental conditions. It is an automatic process, generating areas of interest for further classification. Threat classification requires high resolution images to make use of full spatial information. Classification is a combined human and automatic task, operator aided, in which many different feature extraction processes, such are central moments and key point features, can be performed autonomously. The final threat classification assessment is a task for the human in the loop, optionally supported by automatic behaviour analysis