High-Performance Multi-Platform/Sensor Computing Engine

Over the last decade there has been increasing advances in image processing applied to binoculars and scopes. For instance, hunting and military scopes can allow the user to designate, track, and provide lock indication on a target. However, with advances in neural networks, and associated processors, there are a wide range of additional capabilities that could be incorporated into these optical systems. This Phase I effort will focus on surveying these approaches. Potential offerings include image enhancement, automatic target recognition, semantic segmentation, activity recognition, etc. The Phase I effort will also review how these systems can take advantage of data, both during operation as well as over time. What type of data can be shared among systems to enhance the capability of the operator? How can knowledge be accumulated in order to improve performance over time? Finally, we will review platforms and processor offerings to afford practical implementation. At the end of the Phase I we will create a prototype development plan to be executed in a Phase II. This will include analysis and simulation of expected performance gains against common operating metrics, such as targeting accuracy.