Optimum Dot-Grid Density for Area Estimation with Aerial Photographs

Abstract

Dot-count sampling to estimate area from 1:12,000 scale aerial photographs was tested with grid densities of 4, 16, and 49 dots per square inch. All three densities gave unbiased estimates. Variances may be validly estimated by assuming random sampling and the binomial model when dot density is greater than 27--the optimum density for per-dot reduction in variance. The coefficient of variation can be predicted from dot density of the grid and estimated proportion of dots in the cover type of interest.