Recent studies have shown that the use of a priori knowledge can significantly improve the results of unsupervised classification. However, capturing and formatting such knowledge as constraints is not only very expensive requiring the sustained involvement of an expert but it is also very difficult because some valuable information can be lost when it cannot be encoded as constraints. In this paper, we propose a novel constraint-based clustering approach based on description logics and reasoning for automatically generating constraints from OWL ontology. We apply our approach to classify satellite images. The results have shown that our approach improves the quality of the clustering, while bridging the semantic gap and automating the process of image labeling.