Falls from high altitudes often result into broken bones, usually including the feet. Then often the heel bone, also called calcaneus, is affected. Giving perspectives onto the healing process after taking CT images is the purpose of classifications of these fractures. A fracture classified in a certain category, has a certain probability to heal in regard to the corresponding categories prognosed healing process. So far, the classification of a fracture is heavily depending on the person classifying it. Therefore, a classification algorithm can help to get some consistency into the classification process. This thesis focuses on an automatic classification of calcaneal fractures. We use Sanders classification as our system of classification. To classify a fracture we need to determine the calcaneus. But due to its possible fractures and relocations we need to use the talus as orientation. Throughout this work we apply several different techniques of image segmentation onto the CT images of patients with calcaneal fractures. To get the first classifiable image, we take a look at the sagittal view first. There we need to locate the posterior facet of the talus. To get the posterior facet, we need to use medical knowledge and the shape of the talus. After verifying the calcaneus in coronal view we then are able to classify it.