With the rapid growth of multimodal remote sensing (RS) data, there is an increasing demand for intelligent onboard computing to alleviate the transmission and latency bottlenecks of traditional orbit-to-ground downlinking workflows. While many lightweight AI algorithms have been widely developed and deployed for onboard inference, their limited generalization capability restricts performance under the diverse and dynamic conditions of advanced Earth observation. Recent advances in remote sensing foundation models (RSFMs) offer a promising solution by providing pretrained representations with strong adaptability across diverse tasks and modalities. However, the deployment of RSFMs onboard resource-constrained devices such as nano satellites remains a significant challenge due to strict limitations in memory, energy, computation, and radiation tolerance. To this end, this review proposes the first comprehensive survey of onboard RSFMs deployment, where a unified deployment pipeline including RSFMs development, model compression techniques, and hardware optimization is introduced and surveyed in detail. Available hardware platforms are also discussed and compared, based on which some typical case studies for low Earth orbit (LEO) CubeSats are presented to analyze the feasibility of onboard RSFMs’ deployment. To conclude, this review aims to serve as a practical roadmap for future research on the deployment of RSFMs on edge devices, bridging the gap between the large-scale RSFMs and the resource constraints of spaceborne platforms for onboard computing.

Deep neural networks (DNNs) have shown strong performance in synthetic aperture radar (SAR) image classification. However, their “black-box” nature limits interpretability and poses challenges for robustness, which is critical for sensitive applications such as disaster assessment, environmental monitoring, and agricultural insurance. This study systematically evaluates the adversarial robustness of five representative DNNs (VGG11/16, ResNet18/101, and A-ConvNet) under a variety of attack and defense settings. Using eXplainable AI (XAI) techniques and attribution-based visualizations, we analyze how adversarial perturbations and adversarial training affect model behavior and decision logic. Our results reveal significant robustness differences across architectures, highlight interpretability limitations, and suggest practical guidelines for building more robust SAR classification systems. We also discuss challenges associated with large-scale, multi-class land use and land cover (LULC) classification under adversarial conditions.

Angle estimation of the target in synthetic aperture radar (SAR) imagery plays a crucial role in many aspects, aiming to reverse the target orientation or radar viewing angle. However, many existing regression-based methods suffer from severe angular ambiguity due to geometric symmetry or low image resolution, particularly for unseen targets. To this end, we propose a keypoint-based optimization framework that first estimates the keypoints via a network and subsequently solves the angle. Specifically, we introduce a minimal search-based algorithm to establish keypoints-correspondence, upon which a likelihood function is designed to capture angular ambiguity. To enhance stability, we employ the Kullback-Leibler (KL) divergence to optimize the network over the entire angle space, which is modeled as a Bayesian posterior. The resulting intractable loss term is approximated by Monte Carlo sampling. Additionally, we design a differentiable Levenberg-Marquardt (LM) solver to enable end-to-end backpropagation from the solved angle. Extensive experiments on MSTAR demonstrate that our method outperforms direct regression-based and naive keypoint-based baselines for both seen and unseen targets.