I'm currently trying to analyze rates of shoreline change for an estuarine reserve by extracting & vectorizing shorelines from historic aerial photos, and then comparing those to the modern shorelines.
All of the literature I've read thus far seems to rely on the use of an IR band, which I don't have access to because a large portion of my data are just BW scans of photos taken by the USDA between 1937-1980. Extracting the modern shorelines isn't an issue because I have access to IR via satellite imagery, but I'm unsure how to approach the years in which that isn't an option.
So far, I've tried Iso Cluster Unsupervised Classification in ArcMap, with mixed results mostly depending on the contrast and quality of the images themselves.
Does anyone here know of a more efficient and reliable way to extract shorelines without an IR band?