The work presented provides a feasibility study of the automatic acquisition of the leaf surface and the positions of the trichomes. Furthermore the possibility of tracking trichomes over time was investigated. The raw data are 4D volumetric data (3D + time) of Arabidopsis leaves and trichomes. A robust extraction of the leaf surface and the leaf edge is the first step in this image processing task.                                              Based on these leaf features robust methods (e.g. symmetryanalysis) fit a biologically meaningful reference coordinate system to the leaf. This way trichome positions can be represented in a biologically meaninful manner and all leaves of a time series can be uniformly aligned. Furthermore, it is possibility to find correspondences between two consecutive datasets of a time series.