Characterization of three-dimensional structure using images.

The pair correlation function g(r) and the number density n for particles in a three-dimensional (3D) sample can be determined from a single two-dimensional (2D) image. The 2D image is obtained experimentally with a simple setup: a cross-sectional slab of particles is illuminated with a laser sheet and imaged with a single camera. After image analysis, to find positions of particles in two dimensions, along with their brightness, one obtains g(r), also known as the radial distribution function. The key for attaining high accuracy is to use only the particles that are brighter than a filter level, which we refine to achieve greater accuracy. The density n is obtained from g(r). This method is demonstrated in a dusty plasma experiment. Accuracy is quantified using simulation data; errors of 2% for both the pair correlation function and the number density are achievable. The method is useful for dusty plasmas and colloids.