Multi-level segmentation procedure for measuring the size distribution of nanoparticles in transmission electron microscope images

Electrical, optical, and magnetic properties of synthesized nanoparticles are strongly dependent on their dimensions. The self-assembly of nanoparticles into regular arrays is complicated since the precise control of interactions between particles is required. The development of methods that can rapidly assess dispersed nanoparticle populations is highly desirable. In this paper, we propose a multi-level segmentation procedure for measuring the size distribution and dispersion index of spheroidal nanoparticles from transmission electron microscope (TEM) images of nanoparticles. We focus on the inspection of individual particles in the case of non-overlapping samples, as well as, when particles agglomerate in overlapping samples. Based on the experimental result, the particle segmentation procedure appears to be very effective for the automatic characterization of nanoparticles obtained by the self-assembly process.