Advanced Algorithms for Point Target Detection

Period of Performance: 04/19/2001 - 04/19/2002


Phase 1 SBIR

Recipient Firm

Tanner Research
34 Lexington Avenue Array
Ewing, NJ 08618
Principal Investigator


New developments in neuromorphic concepts permit massively parallel signal processing co-located with the cryogenic focal plane detector array. The Wave Process is a low-contrast moving point target detection algorithm that is efficiently implemented in a compact analog VLSI circuit. It reduces the data rate to downstream sensor processing, while providing data with a higher density of target information than the scene as a whole, resulting in smaller, lighter sensors with improved mission performance. Tanner Research proposes to dramatically increase the Wave Process signal/noise by adding Elementary Motion Detection as a pre-processing function. We will also make major improvements in Wave Process detection of stealth targets by adding Optical Flow and shape-finding post-processing functions. In Phase I, Tanner Research will develop and demonstrate these extensions to the Wave Process algorithm, using an existing high-fidelity simulation and validated imagery data sets representing scenarios relevant to AFRL. We will quantify the improved target detection capability, and establish goals and approaches for the Phase II effort. Compact, low-cost, high performance sensors for point target detection will provide new space-based surveillance capabilities, reliable cruise missile detection, advanced miniature seekers, and collision avoidance for manned and unmanned aircraft. We will pursue technology insertion opportunities with existing sensor platform and seeker primes.