An improved version of the nonlinear iterative projection algorithm is proposed and
applied to the inversion of the light-scattering data for particle sizing. Compared
with the original projection algorithm, the improved one is much more stable,
reliable, and accurate with respect to random noise. The particle-size distributions
retrieved with the improved projection algorithm are independent of the different
starting solutions. The criterion for stopping iteration is fairly clearly determined
and is not a user-dependent parameter of the inversion algorithm. Numerical computer
simulations and experimental results for the mixture of monodispersed size
distributions are presented to allow both the inversion procedure validity and the
instrument performance to be evaluated.
© 2001 Optical Society of America
Equations on this page are rendered with MathJax. Learn more.