Tree trunk detection in urban scenes using a multiscale attention-based deep learning method

• DOI: 10.1016/j.ecoinf.2023.102215
• Publication Date: 2023-07-20T03:06:37Z
• Authors: Rao Li; GuoDong Sun; Sheng Wang; TianZhuzi Tan; Fu Xu
• Keywords: Tree (set theory); Identification

Precise identification of tree trunks contributes to the understanding urban green dynamics. Previous attempts develop trunk detection methods have faced limitations in respect precision and generalization due use hand-engineered features constraint single-species detection. In this study, we construct a new dataset considering object's strong diversity propose deep model detect segment salient or even branches scenes. Comprehensive experiments are performed evaluate our model. The presented method exhibits exceptional performance, evidenced by its outstanding scores across seven evaluation metrics, indicating capability different species, if they exhibit significant variations appearance. Specifically, demonstrates accuracy detecting with intricate furcations, as well effectively identifying that partially occluded.