Precise geographic mapping of remote sensing (RS) data is prerequisite for time series analysis of multiple RS data.
Conventional georegistration of scanned image along satellite track, of which spatial resolutions on earth surface are not
uniform, requires heavy computational efforts. A rapid and precise mapping method, referred to as the Piecewise Affine
Transform (PAT) method, is developed. In PAT method, each polygon of which vertices are defined by ground control
points (GCP) are mapped to target grids, and affine transforms are used in determining the pixel location of scanned
image in mapping data inside of each polygon. The PAT method does not involve any 'proximity searching', and
mapping by PAT method is very rapid. For typical scanned data of a few thousands lines, mapping look up table (LUT)
can be made within a second in personal computer. Using LUT data base, mapping can be easily made by the nearest
neighbour or bilinear interpolation method. The PAT algorithm is tested by mapping MODIS scanned image, and found
precise mapping was achieved.