Utilizing Temporally Invariant Calibration Sites to Classify Multiple Dates and Types of Satellite Imagery

Mapping past time periods (retrospective mapping) using remotely sensed data is hindered by a lack of coincident calibration and validation information. The identification of features of sonic ground cover invariant across time and their use as calibration and validation data addresses this challenge by: (a) streamlining the process of image calibration for multiple dates, and (b) allowing each image to generate its own spectral signature. This study investigates the use of temporally invariant calibration and validation data to map land-cover in Massachusetts, employing five satellite images collected from five separate dates and different sensors. The results indicate that this technique can be used to produce land cover classifications of similar overall map accuracy to published mapping studies. Classification accuracy using this method is highly dependent on the characteristics (radiometric,. spectral, and spatial) of the satellite imagery.