Skip to content

论文arXiv 2611.150669 天前

MiniPose:微缩模型实采数据破解大型工业件6D位姿估计难题 ​

本文由 AI 根据公开内容整理,可能有误,请以原文为准:MiniPose: Miniature-Based Real Data Collection for 6D Pose Estimation of Large Industrial Pipe Segments(arXiv 2611.15066)

核心内容 ​

针对船厂管道等大型重载工业对象6D位姿估计中真值数据采集困难及合成数据Sim-to-Real鸿沟问题,研究提出MiniPose微缩模型实采管线。该方法通过制作等比缩小复制品并使用真实RGB-D相机采集多姿态数据,利用逆缩放变换将深度图转换为全尺度几何一致表示。实验表明,在PVN3D和RDPN6D两种网络中,每类管件仅需混合不超过1000张微缩实图(占比<10%)与10000张合成图,其平均ADD-S AUC指标即显著超越纯合成数据训练基线,证实该方案可作为全尺寸真值采集的有效替代。

技术核心在于保留真实传感器噪声特性同时解决大物体操作难题。系统直接使用微缩模型的原始RGB图像以获取真实外观纹理,对深度图执行简单逆缩放变换以匹配全尺寸几何结构,从而规避了合成渲染难以复现的结构化测量误差。对比测试显示,该管线在两类不同架构的位姿估计器上均表现出泛化增益,且性能逼近全尺寸真值训练效果。这种“微缩实物+合成补充”的数据策略,大幅降低了大型构件感知算法开发中对物理空间与重复搬运的依赖,为工业具身智能提供了低成本数据闭环范式。

关键看点 ​

  • 数据混合策略:每类管件仅需≤1000张微缩实图配合1万张合成图,实图占比不足10%即可提升PVN3D与RDPN6D的ADD-S AUC性能
  • 几何一致性转换:对微缩深度图应用逆缩放变换生成全尺度几何表示,直接复用原始RGB保留真实传感器噪声与外观特征
  • Sim-to-Real弥合:相比纯合成数据,微缩实采数据有效引入了真实深度传感器的结构化测量误差,解决了渲染数据失真问题
  • 全尺寸代理验证:实验证实微缩数据训练效果可逼近全尺寸真值采集,大幅降低大型工业对象感知数据的物理采集门槛

工程与产业观察 ​

MiniPose为大型工业具身感知提供了极具工程价值的数据采集范式。通过微缩模型保留真实传感器噪声与纹理,结合逆缩放几何变换,以极低成本填补了合成数据与全尺寸真值间的鸿沟。这对船舶、石化等难以移动大物体的场景具有直接参考意义,证明了在特定约束下“小比例实物”是解决Sim-to-Real Gap的高效中间态。

英文摘要 (Abstract) ​

6-DOF pose estimation of large and heavy industrial objects, such as large piping systems in shipyards and structural components in petrochemical plants, remains relatively underexplored in robotic perception. Repeatedly repositioning large objects is physically and spatially demanding, making it difficult to acquire real training data with sufficient pose diversity. Synthetic data offer a scalable alternative, but faithfully reproducing real RGB-D observations, particularly the structured measurement errors of real depth sensors, remains difficult. We propose MiniPose, a practical pipeline for miniature-based real RGB-D data collection that addresses both the data-acquisition burden and the sim-to-real gap. We fabricate a scaled-down replica and capture RGB-D data of it in diverse poses using a real RGB-D camera. RGB images are used directly, while depth maps are converted to a geometrically consistent representation of the full-scale object via a simple inverse scaling transformation. This allows training to incorporate real appearance and sensor measurement characteristics that are difficult to reproduce using synthetic data alone. To examine whether the benefit depends on a particular network, we evaluate two RGB-D pose estimators with different formulations, PVN3D and RDPN6D. Mixing at most 1,000 miniature images per pipe type, less than 10% of the training data, with 10,000 synthetic images improves the mean ADD-S AUC across pipe types over synthetic-only training for both estimators. A comparison with training on full-scale real data further suggests that miniature data can serve as a proxy where full-scale collection is difficult, reducing the data-collection burden for large industrial objects.

作者:Hyunjun Kim, Youngjin Seo, Sungwon Seo, Jiyou Shin, Kunwoo Kim, Chanyong Park, Hyungpil Moon

基于本页技术方向关联的整机参数、核心零部件、实测报告与同赛道选型横评。

外部来源

本页事实对应的公开页面,以原文为准。(点击将离开本站)

  1. arXiv 2611.15066:MiniPose: Miniature-Based Real Data Collection for 6D Pose Estimation of Large Industrial Pipe Segments ↗访问 2026-11-14