OrthoRectification parametrisation


Among other options of parametrization, it seems that one can provide the size(in pixel) of the output orthorectified image by providing integer values to ‘outputs.sizex’ and ‘outputs.sizey’. I wonder how it is possible to know / calculate the size of the output image a-priori in this case. Should this be set to the size of the input image? If not, I would appreciate any clarification / link to publication etc.

You should look at the documentation of the outputs.mode parameter:

otbcli_OrthoRectification -help outputs.mode

The parameters ‘outputs.sizex’ and ‘outputs.sizey’ can be tuned, but in several modes, they are actually computed by the application itself. They don’t necessarily match the input dimension, for instance the input image can be rotated w.r.t. the output projection.