Resolution enhancement allows us to assemble a plurality of images into a common image.

Map makers and photogrammeters have previously assembled images into a mosaic in a cut-and paste (euclidean) fashion. Researchers have more recently done this computationally and used the more general affine rather than euclidean model.

Since mosaicing is done mathematically, we are free to choose a still more sophisticated model than the affine model, and the model that we use is the plane projective model. Each image is positioned in terms of its 4 corners (x and y) giving 8 degrees of freedom. Another interpretation of these 8 dof is as follows:

- translation x
- translation y
- dilation x
- dilation y
- shear x
- shear y
- chirp x
- chirp y

An automated algorithm called CEMENT, or cement-mix. Cement-mix was originally developed to combine and mix images that are ``in the same orbit'' (either video or radar, or combinations of both) automatically.

Take a closer look at the CEMENT algorithm,
with some examples of
**video orbits** CEMENTED to stills.

For additional references in the published literature, or other papers, etc., email me:steve@media.mit.edu.