Many problems require finding the coordinate transformation between two images of the same scene or object. One of them is Image Mosaicing. It is important to have a precise description of the coordinate transformation between a pair of images. Image mosaics are collection of overlapping images together with coordinate transformations that relate the different image coordinate systems. By applying the appropriate transformations via a warping operation and merging the overlapping regions of a warped images, it is possible to construct a single image covering the entire visible area of the scene. This merged single image is the motivation for the term ``mosaic''.
Image mosaics allow one to compensate for differences in viewing geometry. Thus they can be used to simplify version tasks by simulating the condition in which the scene is viewed from a fixed position with single camera. Mosaic are therefore quite useful in tasks involving motion or change detection or determining the relative pose of the new images that are acquired. They can be used to determine what parts of the scene visible from that point have been observed.
There are lots of paper about motion parameter estimation about which
can be used in image mosaicing.
A coordinate transformation maps the image coordinates,
to new set of coordinates 
.
Generally, the approach to finding the coordinate transformation relies
on assuming it will take one of the forms in Table 1, and estimating
the two to twelve parameters in the chosen form.
The most common assumption especially in motion estimation for coding
and optical flow, is that the coordinate transformation between frames
is only translation.
Although it is easy to implement, it is very poor to handle large
changes due to camera rotation,panning and tilting.
The other technique is Affine Model which contains translation,
rotation and scale. However, the affine model can not capture camera
pan and tilt and therefore cannot accurately express the seen that we
see in the world.
8-parameter projective model gives the exact eight desired parameters
to account for all the possible camera motions. However, its
parameters have traditionally mathematically and computationally
too hard to find.
Going from first order to second order, gives the 12-parameter
biquadratic model. Increasing the order and number of parameters
doesn't help us too much, because the physical camera model fits exactly
8-parameter projective model.
Therefore, biquadratic model is not suitable for our purposes.
The 8-parameter bilinear model is the most widely-used in the field of
image processing. This model is easily obtained from the biquadratic
model by removing the four 
and 
terms.
Also there is another model very similar 8-parameter bilinear model,
it is 8-parameter pseudo perspective model. It is believed that it also
give very good results to model the physical camera.
In our project We prefer to use 8-parameter bilinear method and
8-parameter pseudo-perspective method in two different approach.
In this project we used two different technique, feature and featureless techniques. In feature based approached, we try to find some matched pair in two images then by using these matched pairs we find the transformation parameters. In featureless approach, we use correlation technique on images.
Table 1: Image coordinate transformations