Journal
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
Volume 14, Issue -, Pages 5421-5435Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JSTARS.2021.3079459
Keywords
Buildings; Semantics; Remote sensing; Spatial resolution; Image segmentation; Convolution; Superresolution; Building extraction; end-to-end network; remote sensing; super-resolution semantic segmentation (SRSS)
Categories
Funding
- National Natural Science Foundation of China [41971280]
- National Key R&D Program of China [2017YFB0504104]
Ask authors/readers for more resources
Higher resolution building mapping is crucial for disaster assessment and emergency rescue, thus a super-resolution semantic segmentation network is proposed to generate higher resolution maps from lower resolution remote sensing images. Experimental results show that the network performs best among different architectures in terms of pixel-level and object-level metrics.
Higher resolution building mapping from lower resolution remote sensing images is in great demand due to the lack of higher resolution data access, especially in the context of disaster assessment. High resolution building layout map is crucial for emergency rescue after the disaster. The emergency response time would be reduced if detailed building footprints were delineated from more easily available low-resolution data. To achieve this goal, we propose a super-resolution semantic segmentation network calledESPC_NASUnet, which consists of a feature super-resolution module and a semantic segmentation module. To the best of authors' knowledge, this is the first work to systematically explore a deep learning-based approach to generate semantic maps with higher spatial resolution fromlower spatial resolution remote sensing images in an end-to-end fashion. The experimental results for two datasets suggest that the proposed network is the best among four different end-to-end architectures in terms of both pixel-level metrics and object-level metrics. In terms of pixel-level F1-score, the improvements are greater than 0.068 and 0.055. Regarding the object-levelF1-score, the disparities between ESPC_NASUnet and other end-to-end methods are more than 0.083 and 0.161 in the two datasets, respectively. Compared with stage-wise methods, our end-to-end network is less impacted by low-resolution input images. Finally, the proposed network produces building semantic maps comparable to those generated by semantic segmentation networks trained with high-resolution images and the ground truth utilizing the two datasets.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available