Accession Number:

ADA539507

Title:

Modeling and Performance Estimation for Airborne Minefield Detection System

Descriptive Note:

Master's thesis

Corporate Author:

MISSOURI UNIV-ROLLA

Personal Author(s):

Report Date:

2008-05-01

Pagination or Media Count:

128.0

Abstract:

Many programs aimed at airborne mine and minefield detection are being pursued and different algorithms are being developed and evaluated to achieve performance specifications. Thus far, no single algorithm or detection architecture has been able to fulfill the performance specifications for different mine and minefield detection scenarios. The reasons for this are numerous. The environment and the operating conditions under which an airborne sensor is expected to perform are highly varied. Also, the performance of airborne sensors and algorithms is highly dependent on the type of targets and environments. Research has been aimed to make the algorithms more robust under these varying conditions, but the studies have been only partially successful. A large amount of data needs to be collected and analyzed to gain insight into detection algorithms and their performance under different operating conditions. Data collection on this scale is time consuming, and costly. Due to this reason, a need exist for a simulation-based approach. One such simulation system is developed and evaluated in this thesis. The factors affecting the performance of an airborne detection system include physical parameters type of background, time of day, data collection parameters swath width, number of steps, in-step and in-flight overlap, and minefield scenarios. Data collection parameters are included in the simulation tool. False alarms and mine statistics are modeled based on the available data collected as a part of the developmental programs. Various mine and minefield detection algorithms are modeled and evaluated. Simulations are run, and Receiver Operating Characteristic ROC curves are used to evaluate the performance at both the mine and minefield levels. Analytical models for minefield detection performance are formulated and used to validate the simulated performance.

Subject Categories:

  • Infrared Detection and Detectors

Distribution Statement:

APPROVED FOR PUBLIC RELEASE