Automated EO/IR Detection Techniques For Floating Objects

We propose to research the application of our novel object detection and classification approach to floating objects using EO/IR sensors. The approach is based on a biologically realistic neural network simulation of the mammalian visual system that has shown promising results in early tests. These results indicate that it will be able to supplement sonar-based systems to reduce false alerts as well as improve detection of mines and other hazards. Further, because the approach is fully trainable, it can be used flexibly in the field to defeat novel tactics. The research in Phase I will demonstrate the feasibility of this approach and provide a Phase II development approach and schedule.