An algorithm has been developped, which makes it possible to automatically extract trajectories of a large number of particles from fast imaging data, allowing a statistical analysis of particles trajectories under various plasma conditions, a better understanding of their influence on plasma properties, and a better characterization of the plasma itself. In this contribution, we focus on results obtained in a radiofrequency parallel plate reactor, where a large amount of micron-sized carbon dust is produced in situ. The use of the rescaled range analysis (R/S analysis) applied to dust particles displacements allows decomposing dust dynamic on different time scales. It is shown that dust displacement is dominated by collisions on short time scales whereas long term behaviour is strongly influenced by large scale plasma fluctuations. Comment: This version contains 8 pages and 5 figures. This work has been presented in the 1st Int. Conf. on Plasma Diagnostics, O8, April 12-16 2010, Pont-\`a-Mousson, France (2010). It has been submitted to Contribution to Plasma Physics
