Integrated Vision System for a Target Seeker

This report summarizes the results of research on An Integrated Vision System for a Target Seeker, (IVS). Detailed explanations of methods, results, simulations, software, etc. can be found in other final reports delivered (A008, A009, A00A, A00B). Figures provided by WL/MNSI on system geometry and...

Full description

Saved in:
Bibliographic Details
Main Authors: Inigo, Rafael M, McVey, E S
Format: Report
Language:English
Subjects:
ARTIFICIAL NEURAL NETWORKS
COMPUTER PROGRAMS
COMPUTER VISION
Cybernetics
GEOMETRY
HOMING DEVICES
ICBM(INTERCONTINENTAL BALLISTIC MISSILE)
IMAGE PROCESSING
IVS(INTEGRATED VISION SYSTEM)
PATTERN RECOGNITION
PIXELS
Target Direction, Range and Position Finding
TARGET RECOGNITION
VERY LARGE SCALE INTEGRATION
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This report summarizes the results of research on An Integrated Vision System for a Target Seeker, (IVS). Detailed explanations of methods, results, simulations, software, etc. can be found in other final reports delivered (A008, A009, A00A, A00B). Figures provided by WL/MNSI on system geometry and requirements are: Field of view of from 100 to 300 micron rad/sec of solid angle. The initial distance between weapon and target is of the order of 100 Km with a closing velocity of 6 to 10 Km/hr. The target dimensions can be between 1m wide-2m long and 3m wide-16m long. Simple geometric considerations show that the above figures are compatible with a feasible system. It is required that the IVS performs in real time and, from the above figures, it is clear that the time from ICBM launching to impact with the intercepting weapon will be just a few seconds. In view of the requirement for real time operation, it was decided that as many parts of the system as possible would be designed using artificial neural networks (ANNs). ANN research has experienced a big increase in activity in the last few years. Among the reasons for this, we can mention the ability of ANNs to recognize patterns even under severe degradation due to noise and other artifacts. In spite of all the research activity on ANNs, progress on their analog VLSI implementation has been very slow. Among the reasons for this are: huge numbers of weights (resistors) and connections (and hence a serious routing problem), heat dissipation, etc.