Toward Cross-Regional Land Cover Mapping Using Regional Consistency and Certainty-Based Active Domain Adaptation With Limited Annotations
Top Cited Papers
Open Access
- 28 January 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 19 (19391404) , 5900-5910
Abstract
Semantic segmentation can automatically classify in remote sensing(RS) images by pixel level into different types of land cover. However, in cross-domain semantic segmentation, how to more accurately identify the boundaries between types of surface features remains a critical challenge to be addressed. In this paper, we introduce a region-aware active selection strategy incorporating region consistency and certainty for RS segmentation. The method, Cross-Regional Active Land Cover Mapping, enhances feature discriminability by maintaining local prediction consistency between pixels and their nearest neighbors in the source image. More accurate identification of boundaries between two types of ground objects positively contributes to improving segmentation accuracy. We conducted four sets of transfer annotation tasks on two representative RS datasets and validated our proposed method through these tasks, and our proposed method has achieved a significant improvement of 1.03% to 38.23% compared to existing methods in the transfer tasks mentioned above and has narrowed the performance gap with fully supervised learning to a minimum of just 0.63%. Our method achieves outstanding performance with minimal annotation effort, offering practical and valuable information for advancing the field of RS.Keywords
Funding Information
- Science and Technology Planning Project of Guangdong Province (2024B0101040005)
- National Natural Science Foundation of China (T2125006)
- National Natural Science Foundation of China (42401415)
- Shenzhen Science and Technology Program (KCXFZ20240903093759004)
- Shenzhen Science and Technology Program (KJZD20230923115106012)
- Fundamental Research Funds for the Universities of Liaoning Province (LJ212510143020)
- Fundamental Research Funds for the Central Universities
This publication has 45 references indexed in Scilit:
- Low Saturation Confidence Distribution-based Test-Time Adaptation for cross-domain remote sensing image classificationInternational Journal of Applied Earth Observation and Geoinformation, 2025
- Open-set domain adaptation for scene classification using multi-adversarial learningISPRS Journal of Photogrammetry and Remote Sensing, 2024
- Self-training guided disentangled adaptation for cross-domain remote sensing image semantic segmentationInternational Journal of Applied Earth Observation and Geoinformation, 2024
- $\mathrm {D^2ADA}$: Dynamic Density-Aware Active Domain Adaptation for Semantic Segmentation2022
- Cross-spatiotemporal land-cover classification from VHR remote sensing images with deep learning based domain adaptationISPRS Journal of Photogrammetry and Remote Sensing, 2022
- Road Segmentation from High-Fidelity Remote Sensing Images using a Context Information Capture NetworkCognitive Computation, 2022
- An attention-fused network for semantic segmentation of very-high-resolution remote sensing imageryISPRS Journal of Photogrammetry and Remote Sensing, 2021
- From local to global: A transfer learning-based approach for mapping poplar plantations at national scale using Sentinel-2ISPRS Journal of Photogrammetry and Remote Sensing, 2021
- Transferable attention networks for adversarial domain adaptationInformation Sciences, 2020
- Adapting Visual Category Models to New Domains2010