Título:	Identification of Leucaena leucocephala in urban landscapes using Street-level images and deep learning.
Autor:	FURUYA, D. E. G. MARRAFON, G. LEMOS, E. L. de FURUYA, M. T. G. GONÇALVES, R. D. S. GONÇALVES, W. N. MARCATO JUNIOR, J. BOLFE, E. L. LIESENBERG, V. OSCO, L. P. RAMOS, A. P. M.
Afiliación:	DANIELLE ELIS GARCIA FURUYA, UNIVERSIDADE DO OESTE PAULISTA; GLEISON MARRAFON, UNIVERSIDADE ESTADUAL PAULISTA "JÚLIO DE MESQUITA FILHO"; EDUARDO LOPES DE LEMOS, UNIVERSIDADE FEDERAL DE MATO GROSSO DO SUL; MICHELLE TAIS GARCIA FURUYA, UNIVERSIDADE DO OESTE PAULISTA; ROBSON DIEGO SILVA GONÇALVES, UNIVERSIDADE DO OESTE PAULISTA; WESLEY NUNES GONÇALVES, UNIVERSIDADE FEDERAL DE MATO GROSSO DO SUL; JOSÉ MARCATO JUNIOR, UNIVERSIDADE FEDERAL DE MATO GROSSO DO SUL; EDSON LUIS BOLFE, CNPTIA, UNIVERSIDADE ESTADUAL DE CAMPINAS; VERALDO LIESENBERG, UNIVERSIDADE DO ESTADO DE SANTA CATARINA; LUCAS PRADO OSCO, UNIVERSIDADE DO OESTE PAULISTA; ANA PAULA MARQUES RAMOS, UNIVERSIDADE ESTADUAL PAULISTA "JÚLIO DE MESQUITA FILHO".
Año:	2026
Referencia:	Urban Science, v. 10, n. 4, 192, Apr. 2026.
Descripción:	Mapping urban tree species supports green infrastructure planning. An essential issue refers to the monitoring of exotic species that may become invasive. Street-level imagery provides a complementary perspective to aerial images for species identification that are difficult to distinguish from above. In this context, our study aimed to evaluate deep learning-based object detection and image segmentation approaches to identify a potentially invasive tree species known as Leucaena leucocephala in an urban environment in Brazil, using 422 street-level images acquired from Google Street View (SV) and mobile phones (MPs). Object detection models (YOLOv8 and DETR) and a foundation segmentation model (SAM, zero-shot) were applied to assess how deep learning paradigms perform under heterogeneous urban imaging conditions. YOLOv8 achieved detection performance with mAP50 above 0.83 and recall up to 0.76. DETR showed domain sensitivity, with mAP50 of 0.45 in SV images and 0.84 in MP imagery. For segmentation, SAM zero-shot achieved 0.92 accuracy and 0.93 F1-score in SV images, decreasing to 0.63 accuracy and 0.66 F1-score in MP images. Overall, this study demonstrates that combining detection and segmentation approaches provides complementary information for urban vegetation monitoring, supporting decision-making related to invasive species management and sustainable urban landscape planning.
NAL Thesaurus:	Urban planning
Palabras clave:	Espécies de árvores invasoras Detecção de espécies de árvores urbanas Segmentação semântica YOLOv8 Detection Transformer Segment Anything Model Planejamento urbano Aprendizado profundo Imagem aérea Invasive tree species Urban tree species detection Semantic segmentation
ISSN:	2413-8851
DOI:	https://doi.org/10.3390/urbansci10040192
Tipo de Material:	Artigo de periódico
Acceso:	openAccess
Aparece en las colecciones:	Artigo em periódico indexado (CNPTIA)