文章来来自于CVPR2019, 文章有三个要点:激光雷达SLAM生成多个点云,多级网络判断点是否存在,无监督学习
首先给上这篇文章的摘要
We propose DeepMapping, a novel registration framework using deep neural networks (DNNs) as auxiliary functions to align multiple point clouds from scratch to a globally consistent frame. We use DNNs to model the highly non-convex mapping process that traditionally involves hand-crafted data association, sensor pose initialization, and global refifinement. Our key novelty is that “training” these DNNs with properly defifined unsupervised losses is equivalent to solving the underlying registration problem, but less sensitive to good initialization than ICP. Our framework contains two DNNs: a localization network that estimates the poses for input point clouds, and a map network that models the scene structure by estimating the occupancy status of global coordinates. This allows us to convert the registration problem to a binary occupancy classifification, which can be solved effificiently using gradientbased optimization. We further show that DeepMapping can be readily extended to address the problem of Lidar SLAM by imposing geometric constraints between consecutive point clouds. Experiments are conducted on both simulated and real datasets. Qualitative and quantitative comparisons demonstrate that DeepMapping often enables more robust and accurate global registration of multiple point clouds than existing techniques.
这是一篇关于怎么将激光雷达生成的多个点云图进行位置旋转偏移后组合到一块,然后根据生成的点融合后来判断这个点是否是真实存在的文章。中间用了两个网络结构,其中第一个是用了L-Net来通过点云的坐标来计算传感器视觉方向向量,然后再根据这个结果对点云来进行坐标的变换,然后将点的信息输入到M-Net当中来判断这个点是否合理存在,即变换后为有效的地图上的点。
在训练上,作者将变换好的坐标整合在一起,把重叠的点[表示变换后一致]拿出来标记为1,表示为点存在,然后无关点拿出来标记为0,因为我们默认在坐标变换后点会重叠在一起,然后根据结果是否符合这个情况来进行参数的更新。
无监督那块,整体来说,就是给出一个在成功条件下,会必然成立的结果,然后根据这个结果设置惩罚,使这个惩罚最小进而更新参数,有点像强化的思路,而这篇文章里对这种惩罚的成立就是以下这个公式。
这块我也是粗读,我只是想了解下作者是怎么实现无监督的,具体大家可以看作者文章P4的3.2,有什么描述不准确的地方,希望大家评论里指出,也是对我的指正,谢谢🙏。
Reference:
Ding, Li , and C. Feng . "DeepMapping: Unsupervised Map Estimation From Multiple Point Clouds." (2018).