Physical Optics Corporation — Department of Defense SBIR Phase I: N172-102

Description

To address the Marine Corps need, Physical Optics Corporation (POC) proposes to develop a new Assistive Compact Convolutional Enhanced Neural Targeting (ACCENT) system based on a new design using convolutional neural network (CNN) software implemented on embedded hardware for assisted target recognition and laser pointing. Specifically, the innovation in image processing algorithms and automatic target designation can be implemented onboard the Future Targeting System (FTS) using a modular, open architecture. As a result, the ACCENT system addresses the fundamental problem of aiding the operator by reducing cognitive load given relatively lower-quality input day and night imagery. This enables the potential weight reduction of FTS by reducing optical design requirements, which directly address Marine Corps requirements. The computationally efficient architecture reduces time to provide feedback thus reducing operator exposure. In Phase I, POC plans to develop concepts for enhancing target detection, identification, recognition, and location capabilities of the FTS. We will perform analysis to show improvements, provide a description of the processing needed, and the development approach to making a prototype system. In Phase II, we will develop a scaled evaluation prototype based on the Phase I results, Phase II performance goals, and Marine Corps requirements for enhanced processing capabilities.