Accession Number : ADA259353


Title :   Multi-Aperture Vision System Coupled to Neural Network Processors


Descriptive Note : Progress rept.,


Corporate Author : FEDERAL ELECTRO-OPTICS INC MEMPHIS TN


Personal Author(s) : Bromley, Keith


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a259353.pdf


Report Date : 15 Dec 1992


Pagination or Media Count : 27


Abstract : As outlined in the original proposal, FEO intends to determine how effectively a neural network coupled to a multiple aperture vision system (MAVS) can be applied to the tracking task. Over the course of the contract FEO will perform seven (7) experimental simulations to evaluate the proposed tracking systems' performance. The October 15, 1992 report discussed the modeling of the first three (3) configurations. The following report discusses the results of the implementation of these first three experiments on the Silicon Graphics IRIS Indigo workstation. Additionally, FEO has completed the optical modeling of experiments four through seven and completed the C++ coding of experiments four and five. FEO has enclosed a DAT tape containing executable code and source code for experiments 1 through 3. The programs can be loaded onto a Silicon Graphics workstation using the tar tape facility. A description of the operation of each program is provided below. Once begun, operation of the executable programs is self-explanatory.


Descriptors :   *NEURAL NETS , *COMPUTER VISION , *AUTOMATIC TRACKING , COMPUTERIZED SIMULATION , ELECTROOPTICS , OPTICAL TRACKING , PERFORMANCE(ENGINEERING) , COMPUTER PROGRAMMING , ARTIFICIAL INTELLIGENCE


Subject Categories : Optical Detection and Detectors
      Target Direction, Range and Position Finding


Distribution Statement : APPROVED FOR PUBLIC RELEASE