Date of Award

2013

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Systems Engineering

First Advisor

Ntuen, Celestine A.

Abstract

The study developed a decision support system known as Visual Analytic Cognitive Model (VACOM) to support the Intelligence Analyst (IA) in satellite information processing task within a Geospatial Intelligence (GEOINT) domain. As a visual analytics, VACOM contains the image processing algorithms, a cognitive network of the IA mental model, and a Bayesian belief model for satellite information processing. A cognitive analysis tool helps to identify eight knowledge levels in a satellite information processing. These are, spatial, prototypical, contextual, temporal, semantic, pragmatic, intentional, and inferential knowledge levels, respectively. A cognitive network was developed for each knowledge level with data input from the subjective questionnaires that probed the analysts’ mental model. VACOM interface was designed to allow the analysts have a transparent view of the processes, including, visualization model, and signal processing model applied to the images, geospatial data representation, and the cognitive network of expert beliefs. VACOM interface allows the user to select a satellite image of interest, select each of the image analysis methods for visualization, and compare ‘ground-truth’ information against the recommendation of VACOM. The interface was designed to enhance perception, cognition, and even comprehension to the multi and complex image analyses by the analysts. A usability analysis on VACOM showed many advantages for the human analysts. These include, reduction in cognitive workload as a result of less information search, the IA can conduct an interactive experiment on each of his/her belief space and guesses, and selection of best image processing algorithms to apply to an image context.

Share

COinS