Combining orientation symmetry and LM cues for the detection of citrus trees in orchards from a digital surface model

This letter proposes a new approach for the automated detection of citrus trees from a single digital surface model (DSM) input. The new approach combines orientation symmetry information and local maxima cues in a probabilistic manner. Experiments are performed on eight test DSMs generated from unmanned aerial vehicle (UAV) images, and the results reveal that the new approach is capable of detecting citrus trees with a high success (overall $F-{1}$-score of 92.5%). The performance of our approach is also compared with leading approaches from the literature and from our own previous work and provided superior or comparable results.