Abstract
Areal-fraction analysis of two-dimensional structures has been studied over the years through areal, lineal, and point sampling. Unbiased estimators of the areal fractions are veil known. Random point sampling has produced the standard variance approximation used in the literature, but systematic point sam-pling requires a different variance approximation due ta the established dependency from point to point. The adjusted vari- ance is derived under the assumption of a general cell model for the construction of the structure under study. In addition, more specific model assumptions result In the joint asymptotic normal-ity of the estimators which also generates an alternative vari-ance approximation.
These variance approximations make use of the available sampling information, so that the usual inferential techniques of interval estimation and hypothesis testing can be employed. Section 9 gives some numerical examples.
Keywords: