Particle number density measurements obtained by image analysis are biased by edge effects and particle overlap. This paper presents an estimator for particle number density that accounts for both edge effects and particle overlap. Using the output of advanced image analysis applied to artificial images, the estimator is shown to enable effective estimation of particle number density for monodisperse particulate systems at high densities. This paper also proposes a single dimension-less number that correlates with the reliability of the inferred PSD based on the likely number of overlaps per particle. As this dimensionless group becomes large, the degree of overlap and occlusion in the images makes image analysis difficult and unreliable. This parameter can aid practitioners in finding the proper sampling conditions to obtain accurate PSD measurements.