>Opps - sorry the new results needed to be switched (as shown below) ...
>
>Results:
>--------
>theta = 1.53766234177657 radians
>x = 3.04923748101702cms
>y = 91.9938099220752 cms
OK, so you already switched to using the upper half of the coordinate plane. Then, assuming we're using the righthanded version of it, the pixel offsets to the right of the image center should yield x>0, and x<0 to the left. If they don't, just change the sign of the pixel offset, either when you're calculating it, or when you use it to calculate theta.
You may start measuring the error, specially the distance as the pixelsize becomes smaller. Move your poles away, measure the physical distance, calculate, see how precise are we.