Abstract
Abstract. This paper addresses the evaluation of algorithms reconstructing a watertight surface from a point cloud acquired on an open scene. The objective is to set a rigorous protocol measuring the quality of the reconstruction and to propose a quality metric that is informative with respect to the various qualities that such an algorithm should have, and in particular its capacity to interpolate and extrapolate accurately. Our approach aims at being more informative and rigorous than previous works on this topic. In addition, we use publicly available data and our implementation is open-source. We argue that a rigorous evaluation of surface reconstruction of open scenes needs to be performed on synthetic data where a perfect continuous ground truth surface is available, so we developed our own LiDAR simulator of which we give a description in the present paper.
Highlights
The topic of surface mesh reconstruction from point cloud has been thoroughly studied judging by the number of approaches recently surveyed by (Berger et al, 2017) and (Khatamian and Arabnia, 2016)
In the context of remote sensing, this topic has first been neglected in favor of 2.5D approaches where the surface reconstruction problem is merely a question of 2D interpolation of possibly sparse height data sampled on a regular grid, leading to the popular Digital Elevation Models (DEMs) or Digital Surface Models (DSMs) used to represent the geometry of the visible surface of a scene seen from above
While mesh distances are sufficient to evaluate the reconstruction of closed, completely scanned objects, we consider that for open scenes, this will only asses the quality of α-interpolation for large α’s, or in other terms the filling of larger holes, where the largest errors are expected
Summary
The topic of surface mesh reconstruction from point cloud has been thoroughly studied judging by the number of approaches recently surveyed by (Berger et al, 2017) and (Khatamian and Arabnia, 2016) The goal of this task is to produce a digital hole-free continuous representation (triangle mesh) of the visible surface of individual objects or entire scenes from sensing data (mostly images and LiDAR scans). A 3D mesh reconstruction algorithm aims at recovering the continuous nature of the underlying scene, such that evaluation metrics need to be based on a continuous representation of the scene that can only be accessed if the data acquisition is simulated on an existing realistic continuous surface.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
