Particle overlap and segregation problems are inherent in on-line coarse particle size measurements using image-based methods. Particle size distributions measured from the exposed layer of a packed particle bed can be seriously biased by these problems and thus, an empirical correction function is commonly employed to transfer measured size data to the actual size data. Although the use of such an empirical correction function might be useful for specific cases, the mechanisms for causing the bias have never been systematically studied. In this regard, studies on particle overlap and segregation problems provide a scientific basis for using image-based methods for particle size measurement and a practical guide for the development of appropriate correction functions. In this research, a Monte Carlo method was used to simulate the phenomenon of particle size percolation at the transfer point of a moving belt conveyor. Important findings from computer simulation were then used for design of the Drop Box Experiment. The results from this research are shown to facilitate the use of image-based methods for on-line coarse particle size measurement.