期刊
EXPERT SYSTEMS WITH APPLICATIONS
卷 202, 期 -, 页码 -出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.eswa.2022.117346
关键词
Building extraction; Deep learning; Pleé iades; Urban
类别
资金
- Istanbul Technical University, Scientific Research Office [MAB-2020-42332]
Automatic building segmentation is a crucial task for various applications, and deep learning-based techniques have been widely adopted. In this study, a new building dataset called the Istanbul dataset was generated for building segmentation. The researchers conducted extensive experiments to explore the ideal architecture and parameters for this task using the Istanbul dataset. The results demonstrated that the Unet++ architecture with the SE-ResNeXt101 encoder pretrained on ImageNet achieved the best performance on the Istanbul dataset.
Automatic building segmentation from satellite images is an important task for various applications such as urban mapping, disaster management and regional planning. With the broader availability of very highresolution satellite images, deep learning-based techniques have been broadly used for remote sensing imagerelated tasks. In this study, we generated a new building dataset, the Istanbul dataset, for the building segmentation task. 150 Ple ' iades image tiles of 1500 x 1500 pixels covering an area of 85 km2 area of Istanbul city were used and approximately 40,000 buildings were labelled, representing different building structures and spatial distribution. We extensively investigated the ideal architecture, encoder and hyperparameter settings for building segmentation tasks using the new Istanbul dataset. More than 60 experiments were conducted by applying state-of-the-art architectures such as U-Net, Unet++, DeepLabv3+, FPN and PSPNet with different pretrained encoders and hyperparameters. Our experiments showed that Unet++ architecture using SE-ResNeXt101 encoder pre-trained with ImageNet provides the best results with 93.8% IoU on the Istanbul dataset. In order to prove our solution's generalizability, the ideal network has also been trained separately on Inria and Massachusetts building segmentation datasets. The networks have produced IoU values of 75.39% and 92.53% on the Inria and Massachusetts datasets, respectively. The results indicate that our ideal network solution settings outperform other methods in terms of building segmentation even without any specific architectural modification. The weights files and inference notebook is available on: https://github.com/TolgaBkm/Istanbul_Dataset.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据