Abstract
Abstract. A generic and practical methodology is presented for 3D surface mesh reconstruction from the terrestrial laser scanner (TLS) derived point clouds. It has two main steps. The first step deals with developing an anisotropic point error model, which is capable of computing the theoretical precisions of 3D coordinates of each individual point in the point cloud. The magnitude and direction of the errors are represented in the form of error ellipsoids. The following second step is focused on the stochastic surface mesh reconstruction. It exploits the previously determined error ellipsoids by computing a point-wise quality measure, which takes into account the semi-diagonal axis length of the error ellipsoid. The points only with the least errors are used in the surface triangulation. The remaining ones are automatically discarded.
Highlights
Terrestrial laser scanners (TLSs) capture the geometry of target object or scene in the form of dense point clouds
signed distance function (SDF) based surface mesh reconstruction is gaining popularity among real – time surface reconstruction works as presented in Newcombe et al (2011), Izadi et al (2011), and Chen et al (2013)
The same object surface is sampled at multiple times each of which are obviously at multiple quality levels. This situation results in data redundancy and the output surface mesh will not be in the uniform quality, if all the high and low quality points are used
Summary
Terrestrial laser scanners (TLSs) capture the geometry of target object or scene in the form of dense point clouds. The most common solution of this redundancy problem is the subsampling the data and generating the 3D surface mesh This solution is not optimal, since the random error characteristic of each individual point is anisotropic. Utilizing of a stochastic surface mesh reconstruction method based on the developed point error model. Kazhdan et al (2006) considered reconstruction as a spatial Poisson problem This approach is improved by the implementation of a parallel computing algorithm (Bolitho et al, 2009). Hoppe et al (1992) developed an approach which first generates a signed distance function (SDF) from the unorganized points using a Euclidean minimal spanning tree, and applies the marching cube algorithm (Lorensen and Cline, 1987) to reconstruct the surface mesh. SDF based surface mesh reconstruction is gaining popularity among real – time surface reconstruction works as presented in Newcombe et al (2011), Izadi et al (2011), and Chen et al (2013)
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
