In solar astrophysics data in form of images enable the observation of surface phenomena in order to investigate underlying processes. Telescopes produce huge amounts of data, which can only be appropriately analyzed via the application of automated image processing. This thesis deals with the implementation of an image segmentation algorithm for the analysis of the solar granulation. The granulation is a distinct feature of the solar photosphere generated by convection. The examined multiple level tracking (MLT) algorithm successfully deals with the problem of solar granulation segmentation by the application of multiple threshold levels. At each level existing structures are extended whereas those adjacent to each other are separated and the results are labeled for further calculations.The segmentation results strongly depend on the quality of the images. The high resolution images of the photosphere, this research work is based on, are provided by the Japanese Hinode Solar Optical Telescope (SOT). In order to improve the implementation of the MLT algorithm, concerning runtime performance, an existing IDL version was re-implemented in C++. A comparison showed that the implementations produce very similar results with an average two dimensional cross correlation of ? = 0.96. The major improvement lies in the enhanced runtime performance. Applied to a Hinode 512 x 512 pixel input image, using three threshold levels, an average speedup by a factor of around 20 could be achieved.In addition to the analysis of observational data the numeric ANTARES model, a 3D radiation hydrodynamics code modeling solar surface convection, was examined in order to determine the structure of the photosphere. Results have been obtained by calculating and comparing correlation height-functions using a model data set. As a verification, results were compared to calculations from former research. These correlations showed no significant deviation and support observational results.