ORSATR-X：面向光学遥感目标识别、基于差分与激励网络的基础模型
ORSATR-X: A Foundation Model based on Differential-and-Excitation Networks for Optical Remote Sensing Object Recognition
把 DINOv3 的通用表征与一个旁路网络结合，用 Weber 局部适配器（中心-周边对比 + 方向梯度）增强低对比度目标的边界判别力，并用自监督蒸馏预训练旁路网络——这套「冻结基础模型 + 轻量旁路增强」的范式可直接迁移到其他低对比度、强背景干扰的视觉检测/分割任务。
原文依据 we pre-train the side network using an efficient self-supervised distillation strategy
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论遥感基础模型（RSFMs）的最新进展已在地球观测（EO）任务中展现出可观潜力。尽管采用自然图像基础模型（如 DINO）为构建 RSFM 提供了一种数据高效的策略，但其强大的泛化能力并不能完全迁移到复杂的遥感（RS）场景，原因在于严重的背景干扰，尤其是在感知低对比度目标等困难目标时更为突出。为此，本文提出 ORSATR-X，一种新型 RSFM，它将 DINOv3 的可泛化表征与一种专门用于激励局部对比度信息的机制有效结合。ORSATR-X 包含两个核心组件：（1）DINOv3 编码器，在有限的遥感预训练数据下提供丰富的特征表征；（2）精心设计的旁路网络，其中包含 Weber 局部适配器（WLA）与多尺度聚合模块（MSAM）。WLA 通过中心-周边对比与方向梯度信息增强，提升复杂场景中低对比度边界的可判别性；MSAM 则通过对多尺度特征的自适应聚合，处理遥感影像中固有的目标尺度变化。此外，我们采用高效的自监督蒸馏策略对旁路网络进行预训练。在场景分类、目标检测与语义分割上的大量实验表明，ORSATR-X 在现有 RSFM 中取得了最优性能，验证了其设计的有效性。
英文原文摘要 1535 字符
Recent advances in Remote Sensing Foundation Models (RSFMs) have demonstrated considerable potential for Earth Observation (EO) tasks. While adopting natural image foundation models (e.g., DINO) provides a data-efficient strategy for building RSFMs, their strong generalization capability does not fully transfer to complex remote sensing (RS) scenarios due to severe background interference, notably in perceiving challenging targets like low-contrast objects. To this end, we propose ORSATR-X, a novel RSFM that effectively integrates the generalizable representations of DINOv3 with a dedicated mechanism for exciting local contrast information. ORSATR-X comprises two core components: (1) a DINOv3 encoder, which provides rich feature representation under limited RS pre-training data, and (2) a carefully designed side network incorporating a Weber Local Adapter (WLA) and a Multi-scale Aggregation Module (MSAM). The WLA enhances discriminability of low-contrast boundaries in complex scenes through center-surround contrast and directional gradient information enhancement, while the MSAM handles inherent object scale variations in RS imagery by adaptive aggregation of features across multiple scales. Furthermore, we pre-train the side network using an efficient self-supervised distillation strategy. Extensive experiments on scene classification, object detection, and semantic segmentation demonstrate that ORSATR-X achieves state-of-the-art performance among existing RSFMs, demonstrating the effectiveness of our design.