Big Ben's Panorama Tutorials

Basic Optimisation

With these we can now pass the information on to PTOptimizer to calculate better values for the input image parameters.... but which parameters do we need to optimise... or more importantly, which don't we need to optimise.

The complete list of variables that can be altered is:

v0 a0 b0 c0 y0 y1 y2 r0 r1 r2 p0 p1 p2 d0 d1 d2 e0 e1 e2

• In this case v0 will need to vary as I know that the axis of rotation was not exactly around the nodal point. 

• a0 b0 c0 may vary but will be best left unchanged from the calibrated values. 

• y0 will be fixed because optimising all yaw angles just wastes time, preventing the panorama from "chasing its own tail"

• r0 and p0 are known (0°) as the first image was carefully levelled.

• The offset (d,e) of the images will vary but these will be optimised at the last stage as a final tweak.

This leaves us with the following v line as our first optimisation.

v v0 y1 r1 p1 y2 r2 p2

The results of this optimisation are then applied to the input parameters and a subsequent optimisation including offset, b and shear is done

v v0 b0 d0 e0 y1 r1 p1 d1 e1 y2 r2 p2 d2 e2 g0 t0

This gives us the following result:

Script

p w1000 h500 f2 v360 u20 n"JPEG g0 q100" k2
m g1 i3
i n"image0.tif"
o f2 y-180 r0 p0 v178.18 a-0.2 b0.292999 c-0.2 d-0.011232 e1.10518 g9.017548 t-8.559522
i n"image1.tif"
o f2 y-55.2707 r-0.716773 p-3.25847 v178.18 a-0.2 b0.292999 c-0.2 d3.277535 e-1.229613 g9.017548 t-8.559522
i n"image2.tif"
o f2 y63.1817 r-2.13526 p-2.13937 v178.18 a-0.2 b0.292999 c-0.2 d-3.921241 e2.249382 g9.017548 t-8.559522

Panorama

Seams

This is a much better result, and good enough at this stage to consider adding the other two images to get rid of the tripod and patch the zenith.  But before we do that let's have a look at the colour/brightness differences between the images.