Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images
Abstract Land Cover and Land Use (LCLU) segmentation plays a fundamental role in various remote sensing applications, including environmental monitoring, urban planning, and disaster management. Traditional models often face limitations in real-time processing and deployment on resource-constrained...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-08-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-07908-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849226307294986240 |
|---|---|
| author | Yahia Said Oumaima Saidani Ali Delham Algarni Mohammad H. Algarni Ayman Flah |
| author_facet | Yahia Said Oumaima Saidani Ali Delham Algarni Mohammad H. Algarni Ayman Flah |
| author_sort | Yahia Said |
| collection | DOAJ |
| description | Abstract Land Cover and Land Use (LCLU) segmentation plays a fundamental role in various remote sensing applications, including environmental monitoring, urban planning, and disaster management. Traditional models often face limitations in real-time processing and deployment on resource-constrained devices due to their high computational requirements. This paper presents a lightweight neural network designed to address these challenges by integrating dense dilated convolutions with pyramid depthwise convolutions for multiscale feature extraction. The proposed encoder-decoder architecture utilizes dense connections to aggregate spatial and contextual information across different resolutions, enhancing segmentation accuracy while minimizing computational overhead. The model’s performance was rigorously evaluated using the NITRDrone and UDD6 datasets, demonstrating a segmentation accuracy of 94.8%, with a significantly reduced parameter count compared to state-of-the-art methods. The compact design of the network facilitates its implementation on low-power devices, enabling real-time LCLU analysis across diverse environmental conditions. This work underscores the potential of lightweight neural networks to advance remote sensing image processing, offering scalable and efficient solutions for practical applications in geospatial analysis. |
| format | Article |
| id | doaj-art-cb7c2233da7d4a7786b736a0debe59ea |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-08-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-cb7c2233da7d4a7786b736a0debe59ea2025-08-24T11:27:33ZengNature PortfolioScientific Reports2045-23222025-08-0115111210.1038/s41598-025-07908-4Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing imagesYahia Said0Oumaima Saidani1Ali Delham Algarni2Mohammad H. Algarni3Ayman Flah4Center for Scientific Research and Entrepreneurship, Northern Border UniversityDepartment of Information Systems, College of Computer and Information Sciences, Princess Nourah Bint Abdulrahman University Computer Science and Artificial Intelligence Department, College of Computing and Information Technology, University of BishaDepartment of Computer Science, Al-Baha UniversityJadara University Research Center, Jadara UniversityAbstract Land Cover and Land Use (LCLU) segmentation plays a fundamental role in various remote sensing applications, including environmental monitoring, urban planning, and disaster management. Traditional models often face limitations in real-time processing and deployment on resource-constrained devices due to their high computational requirements. This paper presents a lightweight neural network designed to address these challenges by integrating dense dilated convolutions with pyramid depthwise convolutions for multiscale feature extraction. The proposed encoder-decoder architecture utilizes dense connections to aggregate spatial and contextual information across different resolutions, enhancing segmentation accuracy while minimizing computational overhead. The model’s performance was rigorously evaluated using the NITRDrone and UDD6 datasets, demonstrating a segmentation accuracy of 94.8%, with a significantly reduced parameter count compared to state-of-the-art methods. The compact design of the network facilitates its implementation on low-power devices, enabling real-time LCLU analysis across diverse environmental conditions. This work underscores the potential of lightweight neural networks to advance remote sensing image processing, offering scalable and efficient solutions for practical applications in geospatial analysis.https://doi.org/10.1038/s41598-025-07908-4Land cover land use LCLURemote sensingDeep learningDense dilated convolutionLightweight implementationContext information aggregation module |
| spellingShingle | Yahia Said Oumaima Saidani Ali Delham Algarni Mohammad H. Algarni Ayman Flah Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images Scientific Reports Land cover land use LCLU Remote sensing Deep learning Dense dilated convolution Lightweight implementation Context information aggregation module |
| title | Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| title_full | Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| title_fullStr | Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| title_full_unstemmed | Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| title_short | Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| title_sort | lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images |
| topic | Land cover land use LCLU Remote sensing Deep learning Dense dilated convolution Lightweight implementation Context information aggregation module |
| url | https://doi.org/10.1038/s41598-025-07908-4 |
| work_keys_str_mv | AT yahiasaid lightweightmultiscaleinformationaggregationnetworkforlandcoverlandusesemanticsegmentationfromremotesensingimages AT oumaimasaidani lightweightmultiscaleinformationaggregationnetworkforlandcoverlandusesemanticsegmentationfromremotesensingimages AT alidelhamalgarni lightweightmultiscaleinformationaggregationnetworkforlandcoverlandusesemanticsegmentationfromremotesensingimages AT mohammadhalgarni lightweightmultiscaleinformationaggregationnetworkforlandcoverlandusesemanticsegmentationfromremotesensingimages AT aymanflah lightweightmultiscaleinformationaggregationnetworkforlandcoverlandusesemanticsegmentationfromremotesensingimages |