Precision rectification of SPOT imagery using the direct linear transformation model

A simple rigorous method for the geometric processing of SPOT images is formulated and evaluated. This method is based on the Direct Linear Transformation [DLTJ model, which is employed affer correcting the image coordinates for systematic distortions caused by Earth rotation and cell size variations due to off-nadir viewing. Corrections for other types of systematic errors are considered through the adjustment. Several experiments are presented using synthetic and real data to evaluate the new method. Simulated data are generated from a general model that imitates the SPOT orbit using the Eulerian parameters, satellite deviations, and velocity vectors, as well as sensor attitude angles as functions of time. The results show that sub-pixel accuracy can be achieved with as few as six control points, if control point errors are kept small. Different parameters that influence the accuracy of the resulting ground coordinates are studied; they include the number and quality of ground control points, image coordinate errors, and base-to-height ratio.