Evaluating geometric accuracy of NeRF reconstructions compared to SLAM method
This work addresses the problem of assessing NeRF's applicability for real-world mapping tasks, but it is incremental as it focuses on a specific evaluation rather than introducing new methods.
The paper evaluated the geometric accuracy of two NeRF reconstructions (using iPhone and robot-sourced data) for estimating the diameter of a PVC cylinder, comparing them to lidar-inertial SLAM in terms of scene noise and metric accuracy.
As Neural Radiance Field (NeRF) implementations become faster, more efficient and accurate, their applicability to real world mapping tasks becomes more accessible. Traditionally, 3D mapping, or scene reconstruction, has relied on expensive LiDAR sensing. Photogrammetry can perform image-based 3D reconstruction but is computationally expensive and requires extremely dense image representation to recover complex geometry and photorealism. NeRFs perform 3D scene reconstruction by training a neural network on sparse image and pose data, achieving superior results to photogrammetry with less input data. This paper presents an evaluation of two NeRF scene reconstructions for the purpose of estimating the diameter of a vertical PVC cylinder. One of these are trained on commodity iPhone data and the other is trained on robot-sourced imagery and poses. This neural-geometry is compared to state-of-the-art lidar-inertial SLAM in terms of scene noise and metric-accuracy.