|Title||The effect of rescaling on fine spatial resolution NDVI data: a test using multi-resolution aircraft data|
|Publication Type||Journal Article|
|Year of Publication||2002|
|Authors||Goodin, DG, Henebry, GM|
|Journal||International Journal of Remote Sensing|
Simulation studies have suggested that pixel aggregation can rescale accurately both Normal Difference Vegetation Index (NDVI) data and fields of biogeophysical values inferred from NDVI. There is, however, a dearth of empirical tests of this hypothesis. This study uses nested NDVI observations acquired from an aircraft-mounted digital camera to assess the spatial properties of measured and rescaled image data. Imagery was collected at four altitudes, yielding spatial resolutions varying from 0.625 m to 3.125 m. The common area from these four images was extracted. The finest spatial resolution image was then rescaled by aggregating pixels to synthesize three rescaled images matching the spatial resolution of the measured images. The spatial properties of these measured and rescaled images were compared, using metrics derived from the variogram. Results demonstrate that the spatial patterns of rescaled data do not correspond to those of the directly observed data. The results suggest that rescaling by averaging blocks of pixels may not adequately preserve spatial properties of very high spatial resolution images and consequent inferred biogeophysical fields.