 This paper presents an original method for detecting and delineating citrus trees using unmanned aerial vehicles based on photogrammetric digital surface models, DSMs. The approach uses an orientation-based radial symmetry transform to handle the symmetry of the trees in the DSM and builds influence regions of each tree to accurately delineate individual canopies. Two efficient strategies are also presented for filtering out erroneously detected canopy regions without height thresholds. Experiments on eight test DSMs reveal that the proposed approach provides superior detection and delineation performances compared to state-of-the-art approaches, supporting a balance between precision and recall measures. This article was authored by Ali Oskine OK and Asli Ostarichi OK.