2026· IEEE Geoscience and Remote Sensing Letters· Vol 23, pp. 6018105-6018105· 0 citations· 20 references
Abstract
Oriented object detection in remote sensing images plays an important role in maritime monitoring, airport surveillance, and traffic management. However, densely distributed small objects and slender-structured objects remain highly challenging to detect because they are susceptible to object adhesion, background interference, and unstable localization. To address these challenges, we propose a cross-scale rotation-aware Mamba state-space network (CRMS-Net) based on oriented R-CNN. The rotation-aware cross-scale enhancement (RACE) module enhances directional geometric representation and cross-level feature interaction, while the multiscale vision Mamba (MSVM) module strengthens multireceptive-field feature extraction and long-range contextual modeling. Experiments on DOTA-v1.0 and HRSC2016 demonstrate that CRMS-Net achieves the mAP values of 82.35% and 98.03%, respectively, outperforming the existing methods.
A hyper look-ahead network is proposed, which incorporates a look-ahead structure (LS), conspicuous feature supplement attention (CFSA), and multiscale feature information process module (MFIPM) in the neck, which outperforms many state-of-the-art object detection methods.
Conventional feature fusion mechanisms largely overlook orientation information, making it difficult to effectively represent objects with diverse rotational patterns. To address this issue, we propose YOLO-RSL, a lightweight rotated object detector that introduces orientation awareness into feature representation, fea...
Jing Zhang, Mas Rina Binti Mustaffa, F. Khalid et al.· International Journal of Adv...· 0 citations
Small object detection in remote sensing images (RSIs) is challenging because imaging degradation weakens object textures, reduces contrast, and blurs boundaries. These effects are further aggravated by hierarchical feature extraction, where repeated downsampling weakens shallow spatial cues before they reach deeper se...
Wei He, Yun-Tao Xu, Qi Qi et al.· IEEE Transactions on Geoscie...· 0 citations
Oriented object detection in remote sensing images is challenged by arbitrary object orientations, large-scale variations, and complex backgrounds. In Oriented R-CNN, direct top-down feature fusion may attenuate local structural cues, while the shared representation in the RoI head may not adequately accommodate the di...
