用预训练视觉模型替代深度传感器，实现低成本高精度单目实时定位与建图。
Dropping the D: RGB-D SLAM Without the Depth Sensor
- 用三个预训练视觉模块替代深度传感器，获取度量尺度的3D特征。
- 在TUM基准上静态序列误差7.4厘米，动态序列1.8厘米，媲美甚至超越主流RGB-D方法。
- 适合追求低成本、低功耗的机器人或移动端实时定位系统开发者。
我们提出DropD-SLAM，一种实时单目SLAM系统，无需依赖深度传感器即可达到RGB-D级别的精度。该系统用三个预训练视觉模块替代主动深度输入：单目度量深度估计器、学习型关键点检测器和实例分割网络。通过膨胀的实例掩码抑制动态物体，静态关键点被赋予预测深度值并反投影至3D空间，形成度量尺度特征。这些特征由未经修改的RGB-D SLAM后端处理以完成跟踪与建图。在TUM RGB-D基准测试中，DropD-SLAM在静态序列上实现7.4厘米均值ATE，动态序列上达1.8厘米，性能匹配或超越当前最优的RGB-D方法，且在单块GPU上运行速度达22 FPS。结果表明，现代预训练视觉模型可作为可靠、实时的度量尺度来源，推动更简单、低成本的SLAM系统发展。
原文摘要 · Abstract (English)
We present DropD-SLAM, a real-time monocular SLAM system that achieves RGB-D-level accuracy without relying on depth sensors. The system replaces active depth input with three pretrained vision modules: a monocular metric depth estimator, a learned keypoint detector, and an instance segmentation network. Dynamic objects are suppressed using dilated instance masks, while static keypoints are assigned predicted depth values and backprojected into 3D to form metrically scaled features. These are processed by an unmodified RGB-D SLAM back end for tracking and mapping. On the TUM RGB-D benchmark, DropD-SLAM attains 7.4 cm mean ATE on static sequences and 1.8 cm on dynamic sequences, matching or surpassing state-of-the-art RGB-D methods while operating at 22 FPS on a single GPU. These results suggest that modern pretrained vision models can replace active depth sensors as reliable, real-time sources of metric scale, marking a step toward simpler and more cost-effective SLAM systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。