A software simulation of a multiaperture vision system (MAVS) that locates a point-source target by means of a neural network processor is presented. Models are developed and tested for one-, three-, and nine-eyelet systems. The advantages and limitations of each MAVS system are discussed. A back-propagation neural network is developed and trained as a target-location processor for each system. The accuracy of the neural network in determining the spatial position of a point-source target placed inside the field of view is then determined. The performance of each of the three systems is compared, and the accuracy of the system is shown to increase with the addition of more eyelets.
Target location by a multiaperture vision system coupled to a back-propagation neural network
Zielortung mit einem Mehrapertur-Erkenungssystem, das mit einem neuronalen 'back-propagation'-Netz als Ortungsprozessor gekoppelt ist
Optical Engineering ; 32 , 11 ; 2703-2708
1993
6 Seiten, 9 Bilder, 6 Quellen
Aufsatz (Zeitschrift)
Englisch
NTRS | 1991
|Signal coherence studies in multiaperture arrays
NTRS | 1972
|Multiaperture overlap-scanning technique for moire metrology [2861-16]
British Library Conference Proceedings | 1996
|Analog Delta-Back-Propagation Neural-Network Circuitry
NTRS | 1990
|