DeepMapping: Unsupervised Map Estimation From Multiple Point Clouds

TL;DR

DeepMapping introduces an unsupervised deep learning framework for aligning multiple point clouds, improving robustness and accuracy over traditional methods by modeling registration as a binary occupancy classification problem.

Contribution

The paper presents a novel unsupervised deep neural network approach for point cloud registration that reduces sensitivity to initialization and extends to Lidar SLAM tasks.

Findings

Outperforms existing registration techniques in robustness and accuracy.

Efficiently solves registration via gradient-based optimization of occupancy classification.

Successfully applied to both simulated and real datasets.

Abstract

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 refinement. Our key novelty is that "training" these DNNs with properly defined 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 classification, which can be solved efficiently using gradient-based 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. Our code is available at https://ai4ce.github.io/DeepMapping/.