Mapping and delineating wetlands of Huntington Wildlife Forest using very high resolution digital color-infrared imagery

The effectiveness of off-site wetland delineation methods using very high resolution digital color-infrared aerial imagery (the color-IR imagery) is compared to the traditional on-site wetland delineation method. The on-site delineation results created using the US Fish and Wildlife Service's National Wetland Inventory (NWI map procedures are compared to the following mapping techniques; heads-up digitizing, hybrid classification, Normalized Difference Vegetation Index (NDVI) and unsupervised classifications (ISODATA) using the same image source. Each of the mapping techniques was applied using the seasonal color-IR imagery. Pair-wise significance tests of the closest mean distances indicated that heads-up digitizing was significantly more accurate than other classification techniques for the color-IR imagery. A combination of the heads-up digitizing and the hybrid classification showed that emergent wetland and scrub-shrub wetlands can be delineated without visiting the ground from the color-IR imagery. Applying logarithmic and hyperbolic sine algorithms to enhance the radiometric property of the color-IR imagery increased delineation accuracy 98% in the spring color-IR imagery and 28% in the fall color-IR imagery. Methods for measuring the accuracy of linear features are reviewed and a new method Points-in-Buffer Analysis (PIBA) is proposed.; Keywords. Wetland boundary delineation, heads-up digitizing, radiometric enhancement, wetland boundary accuracy, point-in-buffer analysis (PIBA)...