A Multi-Objective Semantic Segmentation Algorithm Based on Improved U-Net Networks
Remote Sensing · 30.03.2023 · 15 (7) · 1838
DOI: 10.3390/rs15071838 ↗Kurz gefasst
Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 30.03.2023. Verfasst von Xuejie Hao und 4 weiteren Autor:innen. Die Arbeit wurde bisher 25-mal zitiert (Crossref) und nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Watershed Non-point Source Pollution Prevention and Control Technology and Application Demonstration Project, Ecological protection and restoration of estuarine wetlands in the Yellow River Delta, National Key Research and Development Project.
Abstract (Originalsprache)
The construction of transport facilities plays a pivotal role in enhancing people’s living standards, stimulating economic growth, maintaining social stability and bolstering national security. During the construction of transport facilities, it is essential to identify the distinctive features of a construction area to anticipate the construction process and evaluate the potential risks associated with the project. This paper presents a multi-objective semantic segmentation algorithm based on an improved U-Net network, which can improve the recognition efficiency of various types of features in the construction zone of transportation facilities. The main contributions of this paper are as follows: A multi-class target sample dataset based on UAV remote sensing and construction areas is established. A new virtual data augmentation method based on semantic segmentation of transport facility construction areas is proposed. A semantic segmentation model for the construction regions based on data augmentation and transfer learning is developed and future research directions are given. The results of the study show that the validity of the virtual data augmentation approach has been verified; the semantic segmentation of the transport facility model can semantically segment a wide range of target features. The highest semantic segmentation accuracy of the feature type was 97.56%.
Förderung
- Watershed Non-point Source Pollution Prevention and Control Technology and Application Demonstration Project
- Ecological protection and restoration of estuarine wetlands in the Yellow River Delta
- National Key Research and Development Project
- Natural Science Foundation of China Research
- Fundamental Research
Zitieren
APA
Hao, X., Yin, L., Li, X., Zhang, L., & Yang, R. (2023). A Multi-Objective Semantic Segmentation Algorithm Based on Improved U-Net Networks. Remote Sensing, 15(7), 1838. https://doi.org/10.3390/rs15071838
Vancouver
Hao X, Yin L, Li X, Zhang L, Yang R. A Multi-Objective Semantic Segmentation Algorithm Based on Improved U-Net Networks. Remote Sensing. 2023;15(7):1838. doi:10.3390/rs15071838
BibTeX
@article{Hao2023,
author = {Hao, Xuejie and Yin, Lizeyan and Li, Xiuhong and Zhang, Le and Yang, Rongjin},
title = {A Multi-Objective Semantic Segmentation Algorithm Based on Improved U-Net Networks},
journal = {Remote Sensing},
year = {2023},
volume = {15},
number = {7},
pages = {1838},
doi = {10.3390/rs15071838}
}
EUVIMED