AI-Driven Remote Sensing Image Processing: Algorithms and Advanced Technologies

Remote sensing image processing is a core component of Earth observation, environmental monitoring, agriculture, urban planning, planetary exploration and security applications. Artificial intelligence, particularly deep learning, has transformed the analysis of hyperspectral, synthetic aperture radar (SAR), optical, infrared, subsurface radar sounder and ground-penetrating radar (GPR) data by enabling automated feature extraction, classification, detection, reconstruction and interpretation. However, the performance of current methods can decline under changes in noise, illumination, atmospheric conditions, sensor characteristics, geographic location and acquisition geometry. Heterogeneous data sources also differ substantially in spatial, spectral, temporal and statistical properties, making effective multimodal fusion difficult. In addition, the scarcity of labelled data for specialized applications, the computational cost of large models and the need for reliable edge or onboard deployment remain significant barriers. Recent developments in self-supervised learning, foundation models, lightweight architectures, domain adaptation and advanced fusion strategies offer promising solutions, but their robustness, generalization, interpretability and real-world applicability require further investigation. Progress therefore depends on algorithms that combine accuracy with computational efficiency, adaptability and resilience under realistic remote sensing conditions.This Research Topic aims to advance AI-driven remote sensing image processing through novel algorithms, architectures and technologies that are robust, efficient and adaptable across sensors, environments and applications. It will support research addressing multimodal data integration, low-quality imagery, cross-domain generalization, limited supervision and resource-constrained deployment. Contributions may develop or evaluate lightweight neural networks, semi-supervised and self-supervised methods, super-resolution techniques, domain adaptation strategies, foundation models or task-specific approaches for remote sensing data. Particular emphasis will be placed on studies that connect theoretical AI advances with rigorous experiments and practical applications in Earth observation, environmental monitoring, agriculture, disaster response, urban analysis, planetary science and security.To gather further insights into reliable and scalable AI-based remote sensing image processing, we welcome articles addressing, but not limited to, the following themes:- Hyperspectral image processing, representation learning and classification- SAR image processing, interpretation, detection and change analysis- Optical, infrared, hyperspectral and SAR image fusion- Multimodal and multi-source data fusion for remote sensing- Super-resolution, image restoration, denoising and enhancement- Domain adaptation, domain generalization and cross-scenario learning- Robust learning under noise, illumination variation, atmospheric effects and sensor shifts- Novel neural network architectures for remote sensing image analysis- Lightweight, efficient and compressed models for onboard or edge deployment- Semi-supervised, self-supervised, weakly supervised and few-shot learning- Foundation models, large multimodal models and transfer learning for remote sensing- Subsurface radar sounder and GPR data analysis- Real-world applications in agriculture, environmental monitoring, disaster assessment, urban planning, planetary exploration and security- Interpretability, uncertainty quantification, reproducibility and operational validation Topic editors: Tong Zhou, Hailiang Ye, Elena Donini Source: https://www.frontiersin.org/research-topics/86420/ai-driven-remote-sensing-image-processing-algorithms-and-advanced-technologies