Landcover classification and change detection analysis using high-resolution IKONOS imagery for the Bayview Bog wetland, Ontario

The Bayview Bog wetland in southeastern Ontario is comprised of a complex mix of habitat landcover and species composition. In order to aid in mapping and assessing the ecological health of this important wetland, a satellite remote sensing classification approach was performed using high resolution IKONOS imagery from April 25 2000, and July 14 2006 to provide a tool for wetland evaluation with minimal field reference. This hybrid classification approach integrated unsupervised and supervised classification methods with one meter resolution pansharpend images, image transformation of principal component analysis and tasseled cap transformations with the vegetation indices of Normalized Differential Vegetation Index (NDVI) and Soil Adjusted Vegetation Index (SAVI). This approach yielded highly accurate image classifications for the April 2000 dataset (overall accuracy 84%), while the July 2006 image was less accurate due to cloud and cloud shadow cover (overall accuracy 69%). A limited change detection analysis was performed on a subsection of the study area that identifies landcover conversion that impacts the management of this wetland.