dc.contributor.author | Silva, Rodrigo | |
dc.contributor.author | Marcato Junior, José | |
dc.contributor.author | Almeida, Laisa | |
dc.contributor.author | Gonçalves, Diogo | |
dc.contributor.author | Zamboni, Pedro | |
dc.contributor.author | Fernandes, Vanessa | |
dc.contributor.author | Silva, Jonathan | |
dc.contributor.author | Matsubara, Edson | |
dc.contributor.author | Batista, Edson | |
dc.contributor.author | Ma, Lingfei | |
dc.contributor.author | Li, Jonathan | |
dc.contributor.author | Gonçalves, Wesley | |
dc.date.accessioned | 2022-05-11 20:10:09 (GMT) | |
dc.date.available | 2022-05-11 20:10:09 (GMT) | |
dc.date.issued | 2022-06 | |
dc.identifier.uri | https://doi.org/10.1016/j.jag.2022.102759 | |
dc.identifier.uri | http://hdl.handle.net/10012/18260 | |
dc.description | The final publication is available at Elsevier via https://doi.org/10.1016/j.jag.2022.102759. © 2021. This manuscript version is made available under the CC-BY-NC-ND 4.0 license | en |
dc.description.abstract | Preventive maintenance of power lines, including cutting and pruning of tree branches, is essential to avoid interruptions in the energy supply. Automatic methods can support this risky task and also reduce time consuming. Here, we propose a method in which the orientation and the grasping positions of tree branches are estimated. The proposed method firstly predicts the straight line (representing the tree branch extension) based on a convolutional neural network (CNN). Secondly, a Hough transform is applied to estimate the direction and position of the line. Finally, we estimate the grip point as the pixel point with the highest probability of belonging to the line. We generated a dataset based on internet searches and annotated 1868 images considering challenging scenarios with different tree branch shapes, capture devices, and environmental conditions. Ten-fold cross-validation was adopted, considering 90% for training and 10% for testing. We also assessed the method under corruptions (gaussian and shot) with different severity levels. The experimental analysis showed the effectiveness of the proposed method reporting F1-score of 96.78%. Our method outperformed state-of-the-art Deep Hough Transform (DHT) and Fully Convolutional Line Parsing (F-Clip). | en |
dc.description.sponsorship | This research was funded by CNPq (p: 433783/2018–4, 310517/2020–6, 314902/2018–0, 304052/2019–1 and 303559/2019–5), FUNDECT (p: 59/300. 066/2015, 071/2015) and CAPES PrInt (p: 88881.311850/2018–01). The authors acknowledge the support of the UFMS (Federal University of Mato Grosso do Sul) and CAPES (Finance Code 001). This research was also partially supported by the Emerging Interdisciplinary Project of Central University of Finance and Economics. | en |
dc.language.iso | en | en |
dc.publisher | Elsevier | en |
dc.rights | Attribution-NoDerivatives 4.0 International | * |
dc.rights.uri | http://creativecommons.org/licenses/by-nd/4.0/ | * |
dc.subject | deep learning | en |
dc.subject | convolutional neural networks | en |
dc.subject | line detection | en |
dc.subject | tree branch | en |
dc.title | Line-based deep learning method for tree branch detection from digital images | en |
dc.type | Article | en |
dcterms.bibliographicCitation | Silva, R., Junior, J. M., Almeida, L., Gonçalves, D., Zamboni, P., Fernandes, V., Silva, J., Matsubara, E., Batista, E., Ma, L., Li, J., & Gonçalves, W. (2022). Line-based deep learning method for tree branch detection from digital images. International Journal of Applied Earth Observation and Geoinformation, 110, 102759. https://doi.org/10.1016/j.jag.2022.102759 | en |
uws.contributor.affiliation1 | Faculty of Environment | en |
uws.contributor.affiliation2 | Geography and Environmental Management | en |
uws.typeOfResource | Text | en |
uws.peerReviewStatus | Reviewed | en |
uws.scholarLevel | Faculty | en |
uws.scholarLevel | Graduate | en |