Mean square error (MSE) is used to detect variations in the period
between a pair of self-images formed at two different Talbot planes of
a circular grating (CG) using a beam splitter in a conventional
collimation testing setup. By varying the position of the collimator
with respect to the point source, the collimation state of the input
beam is varied and the computed MSEs are analyzed to deduce the
collimation state. The minimum value of the MSE indicates beam
collimation. For equal sized images, the MSE relates to the sum of the
squared difference between spatially correspondent pixel values of the
images. Since comparison of the spatial information takes place at the
pixels’ level, any small spatial shift between patterns of the two
self-images due to collimation error is detected with precision. The
CG, comprising concentric circular structures, offers added
advantage in terms of error-free alignment, which otherwise is error
prone and cumbersome with widely used linear gratings. It is well
known that self-images formed with circular grating have good fidelity
with less optical distortions and irregularities, especially at
distant Talbot planes. Also, the self-images formed with circular
gratings are less affected by lens aberrations, tilts, misalignments,
etc. Higher sensitivity in beam collimation is achievable, as
self-images of a CG can be recorded at widely separated Talbot planes,
and analyzed using an algorithm which is more responsive toward any
minute difference between them. The suggested method is promising for
a quick collimation setting with good accuracy and enhanced
sensitivity.