Evaluating digital change detection techniques for landscape monitoring in a data-poor tropical environment. A case study of New Providence Island, The Bahamas

Abstract

This thesis addresses the problems associated with conducting remotely sensed change detection for landscape monitoring in tropical island environments, with poor data accessibility and availability. The common unavailability of time-specific in-situ data and the next-to- impossible task of acquiring single sensor imagery with a complete temporal sequence of useable image data for a target location are major obstacles for landscape monitoring in these areas. The overall goal of this study was to evaluate and develop digital change detection techniques incorporating remotely sensed indicators and spatial metrics that can be used with multisensor, multispectral and multiresolution remotely sensed imagery to identify the sequence of landscape changes associated with human disturbance for landscape monitoring, using New Providence Island (The Bahamas) as a case-study.