The freeform surface adaptive interferometer (FSAI) recently has been employed to realize the unknown freeform surface metrology. A near null interferogram should be acquired from the initial interferogram with undistinguished fringes even dark areas. The direct optimization object in FSAI is just the interferogram rather than the focusing intensity characterization in traditional wavefront-sensorless (WFS) adaptive systems. The simulated annealing-hill climbing (SA-HC) mixed algorithm is employed in the FSAI, which has much better convergence than the stochastic parallel gradient descent (SPGD) algorithm, almost as much as the geneticalgorithm (GA). At the same time, it is much faster than GA and thus applicable to the test of volume-produced in the optical shop. Simulations and experiments validating the algorithm feasibility are presented.
Read full abstract