面向野外环境的 DROID-SLAM
DROID-SLAM in the Wild
把逐像素不确定性从多视角特征不一致性中估计出来，并嵌入可微 Bundle Adjustment，为动态场景下的跟踪与建图提供了一种可复用的鲁棒加权机制，可直接迁移到其他需要剔除动态干扰的视觉里程计或重建流程。
原文依据 estimates per-pixel uncertainty by exploiting multi-view visual feature inconsistency
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论本文提出一个鲁棒的实时 RGB SLAM 系统，通过利用可微的不确定性感知光束法平差来处理动态环境。传统 SLAM 方法通常假设场景是静态的，因此在存在运动时会导致跟踪失败。近期的动态 SLAM 方法尝试使用预定义的动态先验或不确定性感知建图来应对这一挑战，但在面对未知的动态物体或几何建图变得不可靠的高度杂乱场景时仍然受限。相比之下，我们的方法通过利用多视角视觉特征的不一致性来估计逐像素不确定性，从而即使在真实世界环境中也能实现鲁棒的跟踪与重建。所提出的系统在杂乱的动态场景中取得了最优的相机位姿与场景几何，同时以约 10 FPS 的速度实时运行。代码与数据集见 https://github.com/MoyangLi00/DROID-W.git。
英文原文摘要 946 字符
We present a robust, real-time RGB SLAM system that handles dynamic environments by leveraging differentiable Uncertainty-aware Bundle Adjustment. Traditional SLAM methods typically assume static scenes, leading to tracking failures in the presence of motion. Recent dynamic SLAM approaches attempt to address this challenge using predefined dynamic priors or uncertainty-aware mapping, but they remain limited when confronted with unknown dynamic objects or highly cluttered scenes where geometric mapping becomes unreliable. In contrast, our method estimates per-pixel uncertainty by exploiting multi-view visual feature inconsistency, enabling robust tracking and reconstruction even in real-world environments. The proposed system achieves state-of-the-art camera poses and scene geometry in cluttered dynamic scenarios while running in real time at around 10 FPS. Code and datasets are available at https://github.com/MoyangLi00/DROID-W.git.