SBIR/STTR Award attributes
We propose two innovative improvements to sensor systems for early detection and tracking. First, develop inexpensive, persistent, infrared optical arrays, using modern optical array sensors and their signal processing systems. The proposed prototype IOA is based on the two operating PANDORA (Persistent And Optically Redundant Array) Tau visible arrays. Second, The Tau tracking optical arrays utilize machine learning (ML) and artificial intelligence (AI) image processing techniques for streak detection and analysis. This capability provides for moving target identification, and enhanced long-range detection at low Signal to Noise Ration (SNR) or in image clutter. The ML, and Convolutional Neural Network (CNN) algorithms replicate processing techniques found in the insect world, where large arrays of very simple detectors are optimized for specialized wide-field imaging tasks. The use of large arrays of simple sensors provides numerous detection and tracking advantages over large, gimbaled, monolithic sensors. Each eye of the fly can detect multiple objects while staring over a hemispherical FOV. Additionally, this imaging sensor approach can degrade “gracefully” – loss of one or a few sensors can be tolerated, since a very high percentage of sensors remain in operation. A key part of our proposed program will be expanding our previous streak detection and image processing program with the Klipsch School of Electric and Computer Engineering at New Mexico State University (NMSU). That program continues to provide new techniques for ML for advanced detection and streak-tracking processing using CNN techniques. Approved for Public Release |21-MDA-10789 (21 Apr 21)
